home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Archive Magazine CD 1995
/
Archive Magazine CD 1995.iso
/
text
/
magazine
/
volume_06
/
issue_03
< prev
next >
Wrap
Text File
|
1995-06-25
|
274KB
|
5,652 lines
Volume 6 • Nº 3 • December 1992
6.3
Positive Feedback
6.3
The comparisons we did last month between Archive and another well-known
Archimedes subscription magazine drew quite a bit of feedback from
readers. Now, as anyone who runs such a publication well knows, people
are usually much quicker to put pen to paper (or finger to keyboard) to
complain than they are to make positive comments. I was particularly
pleased, therefore, that the majority of the feedback was very positive.
6.3
The message that came over loud and clear was, “Don’t increase the size
from A5 to B5” and “If you need to increase the price, feel free − we
think it’s worth it”. We didn’t get a single letter asking us not to
increase the subscription rate. (I can see you’re all kicking yourself
now and wishing you’d written to say, “No increase, please”!)
6.3
So, by popular request... no, that’s not fair... because of the
increased cost of printing and posting the issues that have 80 pages
instead of 64, we are going to increase the subscription from £17 to
£19. However, the increase won’t take effect until 1st February 1993.
Any £17 subscriptions received before then will be honoured even if they
are sent in before they are actually due.
6.3
Negative Feedback
6.3
N.C.S. had a stand at the Computer Shopper Show last month and, quite
frankly, it was very depressing! “Is it a ’486?” was about the only
question some “shoppers” knew how to ask. We told them that it was based
on a “RISC processor” which Acorn had been using in their computers for
nearly six years now. We also told them that IBM and Apple are now
beginning to try using RISC chips in their computers. But then they
asked the price... ...and walked on to the next stand!
6.3
It wasn’t all negative, though. The few who stopped long enough to see
the computers in action were actually quite impressed. The high point
was a conversation with a gentleman who had been working on the Computer
Shopper advice stand. He was an ex-A310 owner about to buy a “486DX
50MHzPQXWZ” (or something!) “because Acorn have been left way behind”.
We showed him Artworks and Impression running on an A540 + CC graphics
card + Eizo T560iT at 1600 × 600. He was absolutely amazed at the speed
− “You’ve restored my faith in Acorn”, he said!
6.3
Wishing you all a very Happy Christmas,
6.3
Products Available
6.3
• A4 Notebooks − At last, we have cleared all the back orders of A4
Model II’s and can now supply them direct from stock. However, the A4
Model I’s, the hard drive-less version, are still on what we call the
RSN list − available ‘real soon now’!
6.3
• Ace colour ProDriver − Ace have produced a version of their ProDriver
for the Hewlett Packard DJ550C colour printer. The price is £47 inc VAT
from Ace or £44 through Archive.
6.3
• Acorn AKF17 monitors − Acorn have some AKF17 standard resolution
monitors that they are trying to get rid of. They are brand new monitors
as supplied with the A3000’s. They don’t have a tilt-and-swivel stand
but otherwise they are functionally similar to the newer AKF30’s and
40’s. The price will be £150 including VAT and carriage from Archive.
6.3
• Acorn Pocket Books − Acorn have changed their policy about Pocket
Books. Non-educational dealers are now going to be allowed to sell them.
In fact, we have several Pocket Books in stock.
6.3
• Acorn Pocket Book accessories − Acorn say that the Pocket Book
accessories will be available at the end of November, i.e. before this
magazine is published. However, I suggest you ring us to check avail
ability before sending orders. The items that we know about are: A-Link
at £46, parallel printer link at £28, 128Kb RAM disc at £74, 256Kb flash
disc at £65, 512Kb flash disc at £111, OPL Editor at £65 and the mains
adaptor at £15. All these are VAT inclusive Archive prices.
6.3
• Arcventure II − The Egyptians − Sherston Software have released their
educational package aimed at 2nd and 3rd year juniors who are studying
Egypt. The children join a simulated archaeological dig to see what they
can find buried beneath the sand. The price is £29.95 +VAT from
Sherston.
6.3
• Automatic embroidery − TechSoft UK Ltd have produced a system which
combines a Poem 500 computer controlled embroidery machine and Archi
medes software called, Arc Embroidery. The system allows images created
as Draw files to be embroidered directly onto a variety of materials
(cotton cloth, felt, denim, jersey, etc) in an area up to 90 × 90mm and
with a variety of user-definable stitches such as satin, ribbed, zig-
zag, etc. The user can also control pattern size, stitch length, fill
density, fill direction, etc. The software even prompts the user for
thread changes as required. The price, including the software, is
£749.95 + VAT.
6.3
• Battery-Pack for BJ10ex − If you are using a BJ10ex with an A4, you
can make the printer portable too by buying a battery-pack costing £46
through Archive. This is trickle-charged from the normal power supply
and allows the printer to be completely free-standing.
6.3
• Budget games − Arcade is a set of four RISC-OS arcade games costing
£5.95 (no VAT) from David Ramsden, 7 Chevet Lane, Wakefield, West
Yorkshire WF2 6HN. Drop Rock is a Repton-type game from Soft Rock
Software for £3.49 (no VAT). Winds of Change is ‘an Archimedes album’
for £1.99 (no VAT) also from Soft Rock Software.
6.3
• Canon BJ200 − Canon have now produced an updated version of their
BJ10ex − the BJ200. This is functionally similar to the BJ10ex but
faster − and it includes a cut sheet feeder as standard. Computer
Concepts are producing a version of their Turbo-Driver software for it.
The price of the software will be £49 +VAT from CC or £53 through
Archive. The printers themselves are available from Archive for £340 inc
VAT.
6.3
• CD-ROM mastering − EESOX offer a low-cost CD-ROM mastering service.
The cost can be as low as £150 for the first master and £50 for
successive copies. These masters can then be used for mass production
later, if required.
6.3
• CD Simulator is a package to simulate the operation of a CD-ROM to
help you in creating and testing your own CD-ROMs. The software costs
£129 +VAT also from EESOX.
6.3
• Colour cards − Two more colour cards are now available. SpeedGraph is
produced by Evolution Trading Ltd in Switzerland (available through MAPS
Ltd in the UK) and costs £350 (DM 950, sFr 900) all inclusive. It
provides 1024 × 768 in 16 colours (chosen from 4096) and up to 800 × 600
in 256 colours (restrictively chosen from 4096) and up to 1024 × 768 in
4096 colours for the display of static pictures. StateMachine G8 Plus is
£249 +VAT (a recently reduced price) offers 1280 × 1024 in 4 colours,
1152 × 900 in 4 colours and 1152 × 848 in 16 colours on monitors that
offer a 64 kHz or greater line rate. For professional work, StateMachine
also offer the G8 Professional which provides 768 × 576 CCIR PAL. This
costs £349 +VAT.
6.3
(At the Computer Shopper Show, CC put one of their cards in my A540. It
gave me, amongst other modes, 1600 × 600 in 16 colours on my Eizo T560iT
with the computer still going at full speed. That will make a huge
difference when it comes to laying out the magazine. I can’t wait to get
a production model. Ed.)
6.3
• Colour driver upgrade − Risc Developments have an upgrade to their
500C driver which will also now work with the DJ550C colour printer. The
price is £15 + VAT + £2 carriage from Risc Developments.
6.3
• Courses galore − Broad Oak Computers have extended their range of
computer courses. They now offer BBC Basic V Programming, C Programming,
Impression II, Pendown, Pipedream, Schema and WIMP Programming in BBC
Basic.
6.3
• Cyborg is an arcade-style, multi-location adventure game in which you
don your cybertronic exo-skeletal suit(!) and face the terrors of an
unknown planet. The price is £25.95 + £1 postage from Alpine Software.
6.3
• Eureka NOT available − We said last month that Eureka was ‘due to
emerge from its long gestation period at the end of October’. This was
based what Longman Logotron’s press and telephone advice. However, they
are now saying that they can ‘see no reason why you should not have it
before Christmas’. We will NOT be mentioning Eureka again until it
really is a ‘Product Available’!
6.3
• Fervour is an action game from Clares. You roll a ball along intricate
pathways with holes to be avoided and bonuses to be picked up en route −
each level must be done within a set time limit. You can even create
your own levels. The price is £24.95 from Clares or £23 through Archive.
6.3
• Guardian (version3) is a virus killer that is ‘capable of killing all
known Archimedes viruses’. This is available free of charge from Digital
Phenomena and is also on this month’s program disc. A technical support
and advice service is also provided free of charge on a Viewdata system
(7E1) on 0705−871531.
6.3
• HP Deskjet550C colour printer − Hewlett Packard have produced an
improved colour printer based on their popular DJ500C. The DJ500C has a
carriage that takes a single cartridge, either a black cartridge or a
colour one. The disadvantage of this is that if you are using the colour
cartridge, blacks come out a muddy colour. (It is possible to get round
this by using colour separation software but it means changing cartri
dges and passing the paper through the printer twice.)
6.3
The new 550C, however, has both a black cartridge and a colour cartridge
on the carriage. With appropriate driver software (such as Ace’s
ProDriver mentioned above), the colour cartridge will automatically be
used for the coloured bits while the black cartridge is used for the
blacks − so the blacks really are black. The Archive price of the 550C
is £590. (The Archive prices of the other HP Deskjet printers have been
reduced. The DJ500 mono is now £380 and the DJ500C is £480.)
6.3
• Impact the Database is Circle Software’s new database which has
mailmerge links to Impression. It also provides direct data import and
export, simple data selection filters, multiple indexing and label
printing. The price is £58.69 inc VAT from Circle Software.
6.3
• IronMask − Turing Tools have produced a password-controlled file
security system for RISC-OS machines. Files shielded by IronMask are
immune from RISC-OS filing operations − they cannot be opened, run,
copied, moved, deleted or renamed. The price is £59 (no VAT) from Turing
Tools.
6.3
• LinkWord is ‘a new concept in language learning’ from Minerva. It is
based on (human) memory techniques to enable you to learn a new language
‘2 to 3 times faster than conventional methods’. It is available in
three languages − French, German and Spanish and costs £39.95 +VAT per
language from Minerva or £44 through Archive.
6.3
• Masterfile 3 − Risc Developments have released a new version of their
Masterfile database so that it is now RISC-OS compliant. The price is
£49 + VAT + £2 carriage from Risc Developments.
6.3
• Monitor stands − We have found a source of monitor stands for A3010
and A3020 computers − in the appropriate colours. They include a slot at
the right for the floppy drive and on the left to enable you to get at
the reset button. They cost £20 each through Archive.
6.3
• NetView is a network utility program that allows an Archimedes station
to view the screen of another BBC, Master or Archimedes on the network.
The price is £34.99 (no VAT) from Sanjay Pattni, Flat 11, 86 Ifield
Road, London SW10 9AD.
6.3
• New titles from G.A.Herdmann Educational are Chemical Spectra at
£19.99 (no VAT) and DTP Images discs 12 and 13 at £8.99 each (no VAT)
which consist of clipart for science teachers.
6.3
• Populous − Krisalis’ latest arcade adventure game for the Archimedes,
Populous, is now available. The price is £29.99 from Krisalis or £28
through Archive.
6.3
• Rhythm-Bed is a full-blown drum sequencer from Clares costing £49.95
or £46 through Archive. Percussion tracks can be built up on screen,
played by the application through the Archimedes’ own sound output or on
a MIDI instrument. The percussion sequence can also be transferred into
other applications such as Rhapsody.
6.3
• RISC-OS 3.10 − If you remember, Acorn’s Managing Director said (or
implied) on 26th October that Acorn would be fulfilling all our back
orders of RISC-OS 3.1 for RISC-OS 2 machines within three weeks. We are
happy to say that we have now had enough to fulfil all back orders.
(However, this does NOT apply to RISC-OS 3.1 for A5000 users − these are
still in short supply.) We have a few left as free stock but it might be
better to ring to enquire about availability before sending in an order
because demand continues to out-strip supply.
6.3
• Sound Engineer is EMR’s application for handling sound samples. It
provides envelope shaping, full use via MIDI, re-pitching, etc. Sounds
can be played through its control panel, via a keyboard window or
through a MIDI instrument. Its features include threshold recording, cut
& paste, auto-looping, 3D FFT printout, direct disc playback, harmonic
wave generation plus spectrum analyser, oscilloscope and amplitude
displays. The cost is £59 +VAT from EMR.
6.3
• TextAid is a set of utilities for manipulating text. It can be used
alongside other RISC-OS packages and provides user-definable macros,
conversion between CSV, TSV and text. It also has the ability to insert
printer control-codes. It costs £9.95 inclusive from Rheingold
Enterprises.
6.3
Review software received...
6.3
We have received review copies of the following: Ancient Greece, Arcade,
Chaos (game), CSVtoText, Cyborg, Desktop Office II, Drop Rock, Easiword
II, Fantasy + Fiddles & Drums (SmArt files), Fervour, Help3, Impact,
NetView, Oh no! More Lemmings, Polyominoes, Precision, Punctuate,
Soapbox, TextAid, The Puddle and the Wardrobe, ThinkLink, Tiles, TV Fun
& Games, Whale Facts, Yes Chancellor II. A
6.3
Government Health Warning − Reading this could seriously affect your
spiritual health.
6.3
“Christmas comes but once a year... ” ...thank goodness! There are more
family break-ups and suicides at Christmas than at any other time of the
year. Why? Well, I suppose that, for the lonely, Christmas is a time
when everyone else is with friends and family − which accentuates their
lonliness. And for those with family problems, having to spend time with
family can be just what they don’t want to do!
6.3
I often think life would be much easier if it weren’t for... people! It
takes such an effort sometimes to build relationships with other people.
Why should we bother? Well, surely it’s because the only truly satisfy
ing things in life come from good relationships with other people.
6.3
The trouble is that developing relationships is SUCH hard work. That’s
true − and what’s more, the Bible tells us why it’s such hard work. It
suggests that unless we get our relationship with God right first, we
really don’t stand much chance of getting our relationships right with
other people. If you want to know how you can get your relationship with
God right, ask a committed Christian − I’m sure they’ll be please to
share the Good News of CHRISTmas with you − the greatest gift of all
time!
6.3
6.3
Norwich Computer Services 96a Vauxhall Street, Norwich, NR2 2SD.
0603−766592 (−764011)
6.3
Abacus Training 29 Okus Grove, Upper Stratton, Swindon, Wilts SN2
6QA.
6.3
Ace Computing 27 Victoria Road, Cambridge CB4 3BW. (0223−322559)
(0223−69180)
6.3
Acorn Computers Ltd Fulbourn
Road, Cherry Hinton, Cambridge CB1 4JN. (0223−254254) (0223−254260)
6.3
Aleph One Ltd The Old Courthouse, Bottisham, Cambridge CB5 9BA.
(0223−811679) (0223−812713)
6.3
Alpine Software P.O.Box 25, Portadown, Craigavon BT63 5UT.
(0762−342510)
6.3
Atomwide Ltd 23 The Greenway, Orpington, Kent BR5 2AY. (0689−838852)
(0689−896088)
6.3
Broad Oak Computers High Street,
Hatfield Broad Oak, Bishop’s Stortford, Herts. CM22 7HE. (0279−507727)
6.3
C.I.S. Unit 2a, 436 Essex Road, London N1 3QP. (071−226−3340)
(071−226−3408)
6.3
Circle Software 33 Restrop View, Purton, Swindon, Wilts SN5 9DG.
(0793−770021)
6.3
Clares Micro Supplies 98 Mid
dlewich Road, Rudheath, Northwich, Cheshire CW9 7DA. (0606−48511)
(0606−48512)
6.3
Colton Software (p5) 2 Signet
Court, Swanns Road, Cambridge CB5 8LA. (0223−311881) (0223−312010)
6.3
Computer Concepts (p40/41) Gaddesden
Place, Hemel Hempstead, Herts HP2 6EX. (0442−63933) (0442−231632)
6.3
Cream PD 38 Baunton, Cirencester Gloucester GL7 7BB.
6.3
Dalriada Data Technology (p31) 145 Albion
Street, Kenilworth, Warkwickshire CV8 2FY. (0926−53901)
6.3
David Pilling P.O.Box 22, Thornton Cleveleys, Blackpool, FY5 1LR.
6.3
Digital Phenomena 104 Manners
Road, Southsea, Hants PO4 0BG.
6.3
EESOX 5 Hillfield Road, Comberton, Cambridge CB3 7DB. (0223−264242)
6.3
Electronic Font Foundry Granville
House, 50-52 Upper Village Road, Ascot SL5 7AQ. (0344−28698)
(0344−872923)
6.3
EMR Ltd 14 Mount Close, Wickford, Essex SS11 8HG. (0702−335747)
6.3
Evolution Trading AG Industries
trasse 12, CH-3178 Bösingen, Switzerland. (010−41 +31−747−6531)
(+31−747−6596)
6.3
G.A.Herdman 43 Saint Johns Drive, Clarborough, Retford, Notts DN22
9NN. (0777−700918) (also 0777−700918)
6.3
Hands-On Midi Software 3 Bell Road,
Cosham, Portsmouth, Hampshire. PO6 3NX
6.3
ICS Ltd (p39) 1 Kington Road, West Kirby, Wirral L48 5ET.
(051−625−1006) (051−625−1007)
6.3
Klein Computers Hasslocherstrasse 73, D-6090 Ruesselsheim, Germany.
(010−49−6142−81131) (81256)
6.3
Krisalis Software Teque House,
Mason’s Yard, Downs Row, Moorgate, Rotherham S60 2HD. (0709−372290)
6.3
Longman-Logotron 124 Cambridge Science Park, Milton Road, Cambridge CB4
4ZS. (0223−425558) (0223−425349)
6.3
LOOKsystems (p31) 47 Goodhale Road, Bowthorpe, Norwich NR5 9AY.
(0603−764114) (0603−764011)
6.3
MAPS Ltd P.O.Box 2841, 87 Vivian Road, Harborne, Birmingham B17 0DL.
(021−428−2696) (021−428−2696)
6.3
Minerva Systems Minerva House, Baring Crescent, Exeter EX1 1TL.
(0392−437756) (0392−421762)
6.3
Morley Electronics Morley
House, Norham Road, North Shields, Tyne & Wear NE29 7TY. (091−257−6355)
(091−257−6373)
6.3
Oak Solutions (p26) Suite 25,
Robin Enterprise Centre, Leeds Road, Idle, Yorkshire BD10 9TE.
(0274−620423) (0274−620419)
6.3
PEP Associates 34 Tiverton Way, Cambridge CB1 3TU. (0223−212251)
6.3
Ray Maidstone 421 Sprowston Road, Norwich NR3 4EH. (0603−400477)
(0603−417447)
6.3
Rheingold Enterprises 17 Ingfield
Terrace, Slaithwaite, Huddersfield HD7 5BJ. (0925−210657)
6.3
Risc Developments (pp6/10) 117 Hatfield
Road, St Albans, Herts AL1 4JS. (0727−40303) (0727−860263)
6.3
Sherston Software Swan Barton,
Sherston, Malmesbury, Wilts. SN16 0LH. (0666−840433) (0666−840048)
6.3
Software 42 109 Ferry Road, Hullbridge, Essex SS5 6EL.
6.3
Soft Rock Software 124 Marissal
Road, Henbury, Bristol BS10 7NP. (0272−761685)
6.3
Spacetech (p32) 21 West Wools, Portland, Dorset DT5 2EA.
(0305−822753) (0305−860483)
6.3
State Machine 75 Upper Wellington Street, Luton, Bedfordshire LU1
5AA. (0582−483377) (0582−480833)
6.3
Techsoft UK Ltd (p9) Old School
Lane, Erryrs, Mold, Clwyd CH7 4DA. (082−43318)
6.3
Turing Tools 149 Campbell Road, Cowley, Oxford OX4 3NX. (0865−775059)
6.3
6.3
Computer Concepts
6.3
From 6.2 page 40
6.3
Computer Concepts
6.3
From 6.2 page 41
6.3
Paul Beverley
6.3
Colton
6.3
New Artwork
6.3
Risc Developments (1)
6.3
New Artwork
6.3
Hints and Tips
6.3
• A3010 joystick ports − several A3010s have been returned to Acorn with
complaints of faulty joystick ports. They are not faulty. The A3010 is
the first Archimedes to have a built-in digital (i.e. Atari-style)
joystick port and because of this, many of the games already available
do not cater for digital joysticks. This is a fairly major problem and
hopefully someone will soon produce some sort of patch (either hardware
or software) which allows digital joysticks to be used in place of the
more common proportional joysticks (such as the Voltmace Delta). Until
then, two sources of information about digital joystick interfaces are
The Leading Edge and Beebug.
6.3
• A4 Portable Guide − on page 31 of this guide, the pin-out diagram for
the AKF18 monitor is slightly incorrect. Pins 13 and 14 (horizontal and
vertical sync) have been reversed.
6.3
• Artworks & Impression dongles − There appears to be a clash between
these two dongles when using an A5000 and a JP-150: printouts appear
with spurious information across them. Computer Concepts say that the
problem only occurs between A5000s and JP-150 printers and add that they
are looking into the problem.
6.3
• Banner-making − I have just seen, in Archive 6.1, a review of a
program called Banner.
6.3
If you own a printer which takes fanfold paper, there is an alternative
way for making banners. (By the way, fanfold paper does not need
sprocket holes if you are careful about the line-up and select friction
feed, so normal paper rolls can be used. For example, the Colourjet 132
users can print banners in this way.)
6.3
The print margins on all the Acorn printer drivers are user-definable
and as long as you want to print out banners shorter in length than A0,
extending the appropriate print margin and setting Draw to A0 landscape
will normally allow satisfactory banners to be produced. (Changing the
print margins is described in the manuals; change only the height of the
page to whatever you wish and leave the width alone.) If print margins
are shown, keeping to these will ensure that none of the picture is
lost.
6.3
The additional flexibility of Draw, especially in conjunction with the
drawfile utilities (!FontFX, !DrawBender, etc) would, in my opinion,
offset most of the ease of use of Banner, except perhaps in primary
schools.
6.3
Note that the printer should not be set to skip perforations, or
problems will result, but I have had no difficulty in printing banners
on an Epson LX and even a Colourjet 132 in this manner.
6.3
For those requiring longer banners, !Draw½ and, I believe, !Draw+
(Careware 13) allow an option to set the work area to the print margins,
which will then enable them to use the longer pages which the printer
driver can handle but, again, for a banner you should set the page to be
much taller than normal in the printer driver and edit the display in
landscape mode. I have used this to produce banners several metres long.
6.3
I am sure Banner is a very useful package and I have not attempted to
use it but I would be unwilling to spend £15 when the same effect seems
to be available from the software supplied with the machine and
Careware. Andrew Garrard, Cambridge.
6.3
• Banner-making again − Recently I needed to print up a large banner for
a meeting, and wondered how I should go about it. Of course, I could
always have bought ‘Banner’ for £15 but, for a once off job, I thought
this just a touch extravagant. This set me thinking... Impression can
define master pages as big as you like... hmm... This is what I did, and
it works!
6.3
1. Open a new document in Impression
6.3
2. Create a new Master Page (Custom Job) − Width: 1000 and Depth: 210
6.3
3. Return to the document and set the Master Page in Alter Chapter
6.3
4. Set the Font Size to 500
6.3
5. Type in “Hello World!” or whatever
6.3
6. Set the printer driver page size to − Height: 1000 and Width: 210
6.3
7. Set the Impression Print Dialogue to print sideways, and fire away
6.3
Warning − this is not ‘printer ribbon friendly’... in fact, it is
positively ‘printer ribbon hostile’! Philip Lardner, Co. Wicklow,
Ireland.
6.3
• CC BJ10ex Turbo Driver − It seems that the only method of using the
Turbo Driver with the PC Emulator is to load !BJ10e and quit it again
before loading the emulator. Chris McIntosh, Blackheath.
6.3
• Programming the Serial Port (Vertical Twist) graphics board − On the
monthly program disc is a series of programs which, in themselves, do
nothing much. However, they show how the 65000 colour modes can be
accessed from Basic and C.
6.3
They mainly show how to draw lines, circles and fill shapes. Most do not
need any interaction. Fill requires five mouse selections to create to
outline − the sixth should be inside the box.
6.3
Lines − just move the mouse and press <select> to change colour.
6.3
15_120ASM converts mode 15/21 sprites to mode 120. Copy the sprite image
to the floppy’s root directory before running. It’s not very fast. Jim
Bailey, Sunderland.
6.3
• Squirrel update − Digital Services have upgraded Squirrel from version
1.07. Currently, Squirrel is being supplied with the original V1.07 disc
and manual and a V1.20 upgrade disc. A new manual is due out soon and
Digital Services will send all existing Squirrel owners this new copy
providing they have registered.
6.3
• VIDC enhancer & RISC-OS 3.10 − There can be problems using very early
Atomwide VIDC enhancer hardware with RISC-OS 3.10. It appears that if
the enhancer is software switched, RISC-OS 3.10 will automatically
switch it on which prevents the computer from starting up correctly. To
solve the problem, you have two choices: (a) at the command line, type
*FX 162,&BC,&80 and switch your machine off, then on again, to check
that the command has worked (full details are on page 14 in the RISC OS
3.10 release note) or (b) return the enhancer hardware to Atomwide with
£5 and they will modify the hardware so that it works correctly.
6.3
• Vision Digitiser. As I concluded in my review of the monochrome
version of this product (Archive 5.12.42), the package is amazing value
for money. However, the supplied software is a little basic. Further
more, Robert Golightly at HCCS tells me that the software module
(promised in the !ReadMe file) to provide * commands and SWIs for users
to write their own programs has a priority below that of the 512-line
upgrade, and is not likely to appear in the near future.
6.3
All is not lost, for the software – at least the monochrome version – is
written using Basic’s in-line assembler, and so can be modified by the
careful user.
6.3
For example, as it stands, the !Vision application always demands 450Kb
of memory. For those with 1M machines, or those who like to multi-task
it with other memory-demanding applications, it would be useful to be
able to reduce this. Examination of the code reveals that a 320Kb screen
buffer is reserved, whatever screen mode you are using. 320Kb will be
needed for mode 21, but only 160Kb for modes 15 and 20, and 80Kb for
mode 12. If you haven’t a multi-sync monitor and want to save 160Kb of
memory when using !Vision, change the following lines of a copy of
!RUNIMAGE, having first checked that the first line of the !Run file
shows that you have Version 1.0, dated March 1992.
6.3
130 DIM M% 210000
6.3
8650 BNE mode_error
6.3
23260 EQUD 163840
6.3
23460 FN_work (163840)
6.3
Then amend line 2 of !Run to read:
6.3
wimpslot -min 288K -max 288K
6.3
I’m working on other possible improvements to the software, such as mode
67 compatibility, but it occurs to me that others might be doing the
same. If you are interested in participating in an informal user-group
for the Vision Digitiser, please send an SAE to Stuart Bell, 23 Ryecroft
Drive, Horsham, West Sussex RH12 2AW. A
6.3
TechSoft
6.3
From 6.2 page 19
6.3
TechSoft
6.3
From 6.2 page 15
6.3
Risc Developments (2)
6.3
New Artwork
6.3
Using RISC-OS 3
6.3
Hugh Eagle
6.3
Many thanks to all those who have sent in material about RISC-OS 3.10.
Hugh has done a lot of work for us in collating all this information, so
read it carefully, check it out on your own systems and let us know if
you find solutions to some of the problems and/or see more problems.
This kind of information is very valuable indeed in terms of helping
other people who have to go through the same learning curve as you, with
the new operating system.
6.3
I think it would also be good to send in a few ‘I hadn’t realised that
you could... ...but it makes life much easier’comments.
6.3
By the way, as at 28/11/92, N.C.S. actually has the £50 upgrades in
stock although A5000 owners will still have to be patient, I’m afraid.
Ed.
6.3
(Throughout the column, references to RISC-OS 3 mean RISC-OS 3.10 unless
otherwise stated. HE)
6.3
After ‘The Long Wait’ and an acute dose of ‘Bent Pin Anxiety’, I had, to
cap it all, ‘The Hunt For The Missing Apps 1 Disc’. After all that, I
was quite surprised to find that the machine actually worked when I
turned it on!
6.3
Reconfiguring it so that it ran as well as it did before the upgrade,
was quite a fiddly process. The main hint I would give to anyone who
hasn’t yet done the upgrade is: watch your Fontmax settings!
6.3
• In the Memory section of !Configure don’t be misled by the illustra
tion on page 94 of the manual. You can not safely leave the “Font cache
limit” setting at zero and expect the operating system to default to a
sensible limit as you can with most of the other memory settings. Unless
you set “Font cache” to a reasonably large amount, you must set a larger
“Font cache limit”. (On a 4Mb machine, I have them set to 128Kb and
512Kb respectively.)
6.3
• If you have a multisync monitor, read the bit on page 119 (in the
section on screen modes!) when deciding what settings to use for anti-
aliasing and cached bitmaps in the Fonts section of !Configure.
6.3
• Keith Lowe adds − Where the upgrade instructions tell you to hold down
the Delete key until the power-on self-test has been completed, they
mean don’t release the key until text has appeared on the screen!
Otherwise your old configuration will still be there and that can cause
no end of confusion!
6.3
• Ed Harris encountered problems in installing the user fit ROM carrier
board from IFEL to his A310. IFEL were very helpful in providing advice
and when Keith was still unable to track down his mistake, they sorted
it out and returned it by return with only a nominal charge to cover
their costs. Many thanks to IFEL.
6.3
• Mark Lambert reports problems trying to install RISC-OS 3 in an A310
fitted with a Watford 4Mb RAM upgrade (the second edition with a
vertical daughterboard holding the processor). Watford have told him
that a simple wire link is required on the RAM board but, as yet, they
have not told him the exact details. Does anyone happen to know the
details of the link required?
6.3
• Steve Hutchinson writes: I was assured that it was quite simple to do
the necessary hardware upgrade to my old A440. In the event, I managed
it without hiccup but it was not quite so simple. I had to remove the
main circuit board from the case − which means unplugging loads of
things in my machine with all its upgrades. If you haven’t reasonable
electronic experience, think twice!
6.3
And is the upgrade worth it? For 1Mb or floppy disc only users, the
answer must surely be yes. It must be a godsend to have the standard
applications, plus fonts, plus modules, all in ROM. For hard disc owners
.... well, what do you think? Some readers have no doubts.
6.3
Ed Harris, for instance, writes: All in all, I have found RISC-OS 3.1 to
be a great improvement on an already excellent operating system (RISC-OS
2) and well worth the cost of the upgrade. However, I wonder whether
others share my disappointment that the improvements were not more
dramatic? Perhaps now is the time for us to begin to compile a Wish List
of things we would like to see in RISC-OS 4.
6.3
Why not send your thoughts either to Archive or to me at the address at
the end of this column?
6.3
Most programs work well in RISC-OS 3
6.3
Robert Macmillan of Colton Software has written to point out (further to
Gerald Fitton’s comments in the Pipeline Column of Archive 6.1) that not
only do all versions of Pipedream from 3.10 onwards work in RISC-OS 3,
they all work better with RISC-OS 3. This is because RISC-OS 3 has the
latest versions of system software built in and these perform better
than the old versions supplied with RISC-OS 2. In particular, the font
manager is much, much faster and the benefits of having the fonts built
in make it possible for floppy disc users to make good use of fonts
without inordinate disc swapping.
6.3
Robert suggests that Archive should publish a table showing which of the
most popular software packages work, and which don’t work or need
upgrading. Judging by the number of needless phone calls he has been
receiving, this would be of great benefit to users considering the
upgrade to RISC-OS 3.
6.3
I think the drawback with this idea is that although most people will
report that most programs work perfectly well (and the list could
potentially fill much of the magazine) some programs that do work well
for most people seem to cause problems for some others.
6.3
However, I do agree that it is important to emphasise that most programs
do work and so, to set the ball rolling, I confirm that, from my limited
experience and that of my friends, the following generally seem to work
well:
6.3
all the standard Acorn applications
6.3
Impression 2.17 (but see below)
6.3
Poster 1.21
6.3
Artworks (all versions)
6.3
Vector
6.3
TWO
6.3
Squirrel
6.3
Pacmania
6.3
E-Type
6.3
Twin World
6.3
Lemmings
6.3
Gods
6.3
Programs that need minor modifications
6.3
• Sigmasheet (Minerva) must have FPEmulator in the same directory, even
though that module is in ROM. (John Waddell)
6.3
• First Word Plus release 1.1 − As my current word processor is First
Word Plus release 1, I was worried by the assertion in the release notes
that this was unuseable under RISC-OS 3.1. This happily proved not to be
the case. If colour 14 within the WIMP palette is redefined, the menus
become legible. I have created an obey file to invoke 1WP (on the
monthly program disc) which loads a modified palette on entry and
restores the default palette on exit. Using this, I have experienced no
problems running 1WP. (Ed Harris)
6.3
• DiscTree does not work correctly under RISC-OS 3. Oak Solutions are
working on the problem.
6.3
• InterWord – My copy of InterWord has its own directory in the hard
disc root directory. It runs in 65Tube, which isn’t where it used to be
under RISC-OS 2, and appears under a different name. To get InterWord
going again, copy the new App2.!65Tube.!RunImage to the InterWord
directory and rename it as 65Tube; it should now be sitting alongside
!IWord in that directory. In the I/WORD.!IWORD.!RUN file, insert the
following line after Set Interword$Dir <Obey$Dir>:
6.3
RMLoad ADFS::HD4.$.IWORD.65Tube
6.3
(where HD4 is the name of the hard disc). Double click on !IWORD in the
usual way to start it.
6.3
A similar method should apply to single floppy machines by substituting
the appropriate floppy name for that of the hard disc above. (Ken Cowap)
6.3
• Holed Out – I have a boot file on the hard disc which sets up the
machine with some applications on the iconbar and some more on the
pinboard. The golf game “Holed Out” finds this environment disagreeable
and won’t start. The following gets it going:
6.3
*CONFIGURE DRIVE 0
6.3
*MOUNT :0
6.3
<shift-break> (machine re-boots to bare-bones iconbar)
6.3
Click Drive 0 icon to get “Holed Out” filer
6.3
Double click !Boot
6.3
Terminate the game with:
6.3
Shift/Break
6.3
*CONFIGURE DRIVE 4
6.3
*MOUNT:4
6.3
<ctrl-break> (machine re-boots to populated iconbar, etc) (Ken Cowap)
6.3
• Chocks Away – The flight simulator Chocks Away isn’t quite so fussy
about a populated iconbar and will generally start normally. However, if
it does object, it may help to try a <ctrl-break> followed immediately
by <escape>, which re-boots to an unpopulated iconbar, then double click
on !Chocks. (Ken Cowap)
6.3
Other program problems
6.3
Mike Williams has supplied a list of programs that he has found don’t
work under RISC-OS 3.1:
6.3
Ground Control Teletext Adaptor (At the A.U. Show, Ground Control
said they weren’t aware of any problems. HE.)
6.3
Clares’ Arcade Games
6.3
Quazer
6.3
StarTrader
6.3
Arcendium
6.3
Corruption
6.3
Terramex
6.3
Boogie Buggy
6.3
Ed Harris has found several public domain applications that have failed
to run:
6.3
Mandelplot (garbage on screen − machine locks)
6.3
ChartDraw (“illegal window handle” error)
6.3
Hyper (“illegal window handle” error)
6.3
• Computer Concepts admit that ShowPage does not work with RISC-OS 3 and
say that they have no plans to get it working with RISC-OS 3. There just
is not the demand for this product to make it worthwhile doing the
upgrade which is more like a complete re-write for the new operating
system.
6.3
Denis Fox reports problems with:
6.3
• Impression (2.17) and LaserDirect (2.05) work much more slowly, even
with Fontsize set to 256Kb and Fontmax to 512Kb. There are regular poor
printouts, sometimes with patterned effects in what should be black
areas, hang ups when transferring lots of text between documents and
continual error messages when printing. (I have a similar setup but have
not noticed any deterioration in performance. HE.)
6.3
To help the speed, I suggest you try NOT using the ROM fonts − then you
can have Quick Text switched on. To do this, add a FontRemove
Resources:$.Fonts. in the !Fonts.!Boot file before the FontInstall
<Obey$Dir>. command.
6.3
I have once had the patterned effect in black areas but quitting and
restarting the printer driver cleared it. Ed
6.3
• Multistore V11 not printing with a Star LC24-200. (At the AU Show,
Minerva said there may be problems with earlier versions and that users
experiencing such difficulties should contact them quoting their version
number. HE.)
6.3
• Scanlight Plus 256 not working at all. (There are no known problems.
Are you sure you have a power supply connected to the board? You do need
the p.s.u. connected even on the A5000. Ed.)
6.3
• A number of people have had problems (mainly “illegal window handle”
errors) with versions up to 1.07 of my PD application PickAPic. I found
that there was an error in the code that only occasionally caused
problems in RISC-OS 2, but caused the program to crash quite often in
RISC-OS 3, presumably because of a change in the way the window manager
numbers its windows. Anyone who wants an upgrade to the latest version
please write to me at the address at the bottom of this column. (Hugh
Eagle)
6.3
• Droom – I can’t get Droom to work. It reports “Not enough system
sprite memory” whether I use Memalloc and the *Spritesize command (this
used to work in RISC-OS 2), or the new *ChangeDynamicArea command, or
whether I set the System sprites slider in the Task Manager window.
(Hugh Eagle)
6.3
• Compression − Computer Concepts have informed me that the release note
for version 3.10 is slightly too pessimistic when it claims that
Compression (up to version 1.10) does not work on RISC_OS 3.10. What
apparently happens is that sometimes when you compress a directory with
several files, the last file, instead of being processed, produces an
error message. It can then be compressed singly. (In the first two weeks
of working with the new OS, I have had no trouble with Compression, so
the problem appears to be marginal. HE) (Jochen Konietzko)
6.3
• Impression and the Pinboard − Can anyone tell me why the Pinboard has
such problems with Impression documents?
6.3
When I drag a compressed document onto the desktop, the next reset will
produce an error message “Sprite does not exist”. Without Compression,
the document does appear on the desktop but the Impression icon is
replaced by the standard application icon. (Jochen Konietzko) (I seem to
have no difficulty with Impression icons on the Pinboard in my
Compression-less setup. HE)
6.3
• PC Emulator (version 1.6) – I applied the patch to my copy of the PC
Emulator (Version 1.6) and changed the filetype of the DOS partitions to
DOSDisc as recommended in the release notes. I then found that the
machine locked whenever I ran the emulator, even after a <ctrl-break>.
Things worked perfectly, however, after the machine had been powered off
and restarted. (Ed Harris)
6.3
• Trouble with !FontFX − It has already been documented, at least twice,
that !FontFX will not run under RISC-OS 3 and the suggested “fix” is to
reset Font$Path using *SetMacro. Having done this, I still can’t get it
to use the Trinity and Homerton fonts supplied in ROM. To get round the
problem, I have renamed the original RISC-OS 2 versions OldTrinity and
OldHomertn and installed them in my !Fonts directory on disc. Now
choosing these from the fonts list in !FontFX works beautifully. So far
there have been no side effects!
6.3
It occurs to me that the problem may be that !FontFX is looking for
files called IntMetric and Outlines (there are lots of references to
these names in the !RunImage) whereas the files in ResourcesFS are
stored as IntMetric() and Outlines(). The real answer would be for the
program to be amended to accept either version of the filenames. I shall
speak to DataStore to find out if they intend to make the necessary
alterations. (Tom Harkness)
6.3
The release notes say that FontFX must be more recent than 4.31 to work.
My version, 4.08, works fine so long as you don’t use ROM fonts. If you
do try to, it reports ‘Can’t find outlines’ and promptly exits from the
iconbar! All other fonts seem fine, even from multiple directories.
(Steve Hutchinson)
6.3
• Ovation – Ovation meets a problem or two with RISC-OS 3 despite its
recent update (version 1.35S). If text is indented (e.g. at the start of
a paragraph) and is also justified, the correct format is presented on
screen but when printed, the indented line projects to the right beyond
the text frame. It is understood that Risc Developments are working on
this. Also, if Ovation encounters a faulty outline font (e.g. containing
a zero-width character, not uncommon in some PD fonts) it puts out an
incorrect error message saying that the font cache is full. This bug is
more likely to be purely Ovation’s rather than in conjunction with RISC-
OS 3, but hasn’t been resolved as yet (except of course by ensuring that
outline fonts are fault-free). (Ken Cowap)
6.3
• ArcDFS – Since fitting the RISC-OS 3 upgrade, ArcDFS no longer works
correctly and consistently comes up with error messages. This causes me
great problems since, throughout our group of companies, we possess some
50+ Acorn BBC units ranging from the old BBC Model B’s (yes, they’re
still going strong after eight years of service) through to Master
128’s.
6.3
We are only just moving on to Archimedes and to maintain our wealth of
information accumulated on our Beebs, I am transferring files from DFS
to ADFS via a set of external 5¼“ disc drives I have had connected to my
A3000. This has worked perfectly in the past with complete surfaces of
DFS discs being transferred to ADFS in seconds. Without ArcDFS, I am
left with only one solution − to *MOVE individual DFS files to 640Kb
ADFS discs using a Master Series Computer and then transferring these
640Kb discs to the Archimedes. Clearly, this is a retrograde step after
the convenience of ArcDFS.
6.3
I have been in contact with Dabs Press and, although they are aware of
the problem, they could not tell me whether or not they were going to
make the necessary modifications to their software! I sincerely hope
they receive a large enough response from their users to persuade them
to produce an upgrade for this product. (S. J. Furnell)
6.3
• !Alarm – A couple of tricky little bugs exist in the new Alarm
application. I’ve already reported these to Risc User and trust they
will advise Acorn in turn. If the Repeating Alarm facility is used to
operate, say, every first Monday of every month, it malfunctions when
the next month comes up to December. It advances a year instead of a
month, going from November 1992 to December 1993.
6.3
Perhaps less important, but still a nuisance if you want to use it, is
the bug in the automatic BST to GMT facility. It refuses to accept
entries for two different years such as were needed at the recent change
on 25th October. Before that date, it should have been in order to enter
6.3
BST starts on Sunday 28th March 1993 at 01:00
6.3
BST ends on Sunday 25th October 1992 at 01:00
6.3
but Alarm invalidates such an entry by altering the “starts” date to be
earlier than the “ends” date. If one capitulates and enters the two real
changeover dates for 1992, (hoping that Alarm will calculate forward
into 1993) Alarm does indeed respond with a date in 1993 but only about
11 months ahead, on 1st March 1993. There is a fundamental difficulty
for the programmer here. The changeover date doesn’t, as far as I know,
follow a properly defined rule and tends to be arbitrary. I suspect that
the programmer intended to issue a reminder to the operator 11 months
after the last change so that the true date, known by then, could be
entered. However, the algorithm obviously needs re-examination. (Ken
Cowap)
6.3
Can readers (or the manufacturers) offer any advice on how to overcome
problems with these programs (or any others that have proved
troublesome)?
6.3
Hardware problems
6.3
• Oak SCSI “Free Space” display – My Oak SCSI hard disc won’t use the
new ‘Free Space’ window but uses the old RISC-OS 2 ‘stop-everything-and
press-space’ window. The Release notes suggest that SCSIFiler will fix
this for the Acorn SCSI card, but it doesn’t appear to for my Oak
(v.1.16). Any thoughts? (Steve Hutchinson) (The solution to this was
explained in Archive 5.5 p40. Ed.)
6.3
• Digitisers, SCSI and the RS232 port – I managed to get RISC-OS 3
fitted to my faithful A440 only to discover that my Technomatic
(Brainsoft) Multipod Professional digitiser not only does not work with
RISC-OS 3.10 but the computer will not work with the podule fitted.
Richard Brain says that he was not aware of a problem and there are
currently no plans for an upgrade, but he will have a look when he gets
time to see if an upgrade is possible. In the meantime, I have no video
digitiser, no stereo digitiser and no extra RS232 ports.
6.3
I mentioned the problem to Malcolm Banthorpe and he tells me that his
Pineapple digitiser is now “unwell” with RISC-OS 3.10. I wonder whether
there are some differences in the way that podules download their ROM
modules and initialise under RISC-OS 3.10?
6.3
I also hear rumours that people are experiencing interrupt problems with
SCSI interfaces causing data loss on the RS232 port (interrupts not
returning quickly enough to avoid loss of incoming data on the port).
Apparently, uncached Morley cards may be particularly prone but I have
no first hand experience. It certainly fits in with Malcolm Banthorpe’s
experience on his digitiser which seems to be suffering from an overdose
of interrupts. (Tim Nicholson)
6.3
Laser Direct RISC-OS 3 printer driver?
6.3
It seems barely credible, but a whole year after RISC-OS 3 was first
released, the Rolls Royce of Archimedes printers still lacks a RISC-OS 3
printer driver. For many purposes this doesn’t matter, but rotated fonts
seem a bit pointless if you can’t print them out, and users with more
than one printer would like to be able to access their LD using the
Print Manager.
6.3
Readers have been getting contradictory explanations from Computer
Concepts: one was told that they’ve been too busy writing Artworks,
while another was told that Acorn haven’t given them a printer driver
specification yet! (There’s an element of truth in both. Let’s just say
that it is a non-trivial task to write a printer driver to work with
Acorn’s Print Manager. Ed.)
6.3
Other printer problems
6.3
• Epson MX100-III – My printer is an ageing Epson MX100-III. Using the
standard settings provided by the Acorn printer driver, I find thin
lines can disappear. e.g. the staves in a Rhapsody score. Using the
highest resolution setting (120 x 216 dpi dithered) produces an
excellent standard of output but 1.5 to 2 times slower than the RISC-OS
2 driver (at a resolution of 120 × 144 dpi). (Ed Harris)
6.3
• Print buffering – I have an Archimedes 420/1 expanded to 4Mb memory,
with twin Cumana 5¼“ disc drives and recently equipped with ARM3 and
RISC-OS 3.1. Ever since I installed RISC-OS 3.1, I have had a problem
with printing. I use Impression most of the time but the same problem
arises with Draw, Acorn DTP, DrawPlus, etc, etc. I shall put the problem
in the form of a question: What is the point of a printer buffer that
does not release control of the machine back to the user until the print
job is finished?
6.3
I wrote to Acorn as follows: Whatever setting I use for PrinterBuffer
Size, it just does not seem to make any difference. I have 4Mb of memory
and use Impression a great deal. Sometimes I set the PrinterBufferSize
as high as 512Kb but the computer does not return control to the desktop
until the document is completely printed. Under RISC-OS 2, I used a
printer buffer program and found that I was able to work on a new
document within a minute or two of sending the previous one to the
printer provided the buffer was set high enough to accommodate all of
the data being sent to the printer. Now the machine is always tied up
until the printing is complete. How does one get the printer buffer to
work? I use *Configure PrinterBufferSize to set the size of the buffer.
Is there another command that I can use to make it operational and is it
possible to turn it off (*Configure PrinterBufferSize 0Kb sets it to the
default of 1Kb) and use one of the many printer buffer programs that are
available?
6.3
Acorn (Peter Dunn) replied: “Although the printer buffer does work in
that data is passed into the buffer, control is not released back to the
user until the print job is finished.”
6.3
I have tried RMKilling the BufferManager but that stops all data being
sent to the printer.
6.3
I have devised a work-around but it is a little long-winded:
6.3
1. Have two versions of the chosen printer driver on the iconbar and
make them active.
6.3
2. Set one (F) to send the data for the printer to a file on the hard
disc and set the other (P) to send the data to the parallel or serial
printer.
6.3
3. Select (F) by clicking on it and ‘print’ the document. A file called
Printout will appear on the hard disc.
6.3
4. Change the name of the file called Printout to, say, Printoutx.
6.3
5. Select (P) by clicking on it and drag Printoutx onto the Printer icon
that is yellow.
6.3
The file will now print in the background and control is instantly
returned to the user. Step 4 is not strictly necessary but is useful if
you are printing longish documents and wish to, as it were, stack them
ready for sending to the printer. If one does not change the name to
something unique, there is a danger of overprinting one file with
another. Instead of the hard disc one may, of course, use a floppy or
RAMFS.
6.3
This is quite a laborious work-around when one has a printer buffer
built in (so one would think) to RISC-OS 3.1. A better arrangement must
be possible and so I should be grateful for any hints or tips the
readers can give. (M. P. Ireland)
6.3
• Epson FX-105 – The driver for the Epson FX-85 does not work with the
Epson FX-105 even though the printer manual suggests that they are
virtually identical apart from the carriage width. The printout is
always aligned along the left hand side even when it is supposed to be
centered, etc. Furthermore, large letters, being sent in several
separate lines, are often incorrectly aligned from top to bottom. When I
raised this in my letter to Acorn, they replied as follows:
6.3
“The Epson FX-105 does not support the ESC $ (remove leading white
space) options supported by the FX-85 but does however support NLQ text
printing. The decision by Epson not to support ESC $ in its newer
printers has only recently come to light when testing the Epson LX-100.
We have written a driver for the LX-100 which I am sure you will find
works correctly with your printer.”
6.3
A copy was sent to me and I can confirm that it does indeed work. (M. P.
Ireland)
6.3
• Bug in !Printers − If you configure two printers of the same type
(e.g. a 300dpi HPLJ-2 and a 150dpi HPLJ-2) then delete the first one,
!Printers crashes with an address error. (Mike Williams)
6.3
• LaserJetII printer driver and Qume Crystalprint Series II – My problem
concerns the LaserJet II printer driver. Quite simply, I always get an
extra blank sheet for every page printed. Also stray characters appear
in the left margin. The Qume CrystalPrint Series II printer is very
similar to that used initially by Computer Concepts for their Laser
Direct series, but mine isn’t a Laser Direct model.
6.3
In the early days of RISC-OS 2, I vaguely remember I had a similar
problem that was solved only when Acorn (fairly quickly) brought out
another printer driver.
6.3
In my efforts to solve the problem, I’ve tried lots of things – like
changing margins, etc, etc. I’ve had long discussions with Gerald
Fitton, too. From him, I know that others have plenty of problems,
though perhaps not quite like mine. The silly thing is that all was OK
with the RISC-OS 2 driver but, despite the impression given in Risc
User, neither of us thinks that RISC-OS 2 drivers can be used with RISC-
OS 3. Is this correct, I wonder? If they can be used, that would be a
good temporary solution. (John Jordan)
6.3
• Getting the page length right – After quite a number of abortive
attempts to set the page length for 12 inch fanfold paper, I finally
found out that any changes I make become effective (at least on my
combination of machines) only after I have saved all changes, quit and
then reloaded the printer driver. (By the way, the correct setting for
my Fujitsu DL1100 printer, using the Epson LQ860 emulation is, for some
inscrutable reason, 12.309 inches, 72 lines). (Jochen Konietzko)
6.3
• Panasonic KX-P1124 – Probably my most difficult problem was with the
printer drivers. The old RISC-OS 2 drivers just have to be abandoned −
they merely generate an error message. My printer happens to be a
Panasonic KX-P1124 (24-pin dot matrix, monochrome) but Acorn haven’t
seen fit to include any Panasonic printers at all. Though they blandly
say that one of the others will probably be near enough, in fact, it
requires a considerable amount of research to be confident that the
range of control codes is (a) sufficient to fully exploit the machine’s
capabilities and (b) doesn’t contain any alien codes to cause unwanted
results. At the end of the day (or several!) I reduced the probables to
two, the Epson LQ860 and the NEC P6+ (despite the fact that both
machines are colour printers). I had misgivings about the NEC which
includes, in its control code repertoire, some which started with 28
instead of the old familiar 27 (Escape). The matter was finally clinched
when I was able to question an Acorn representative at the Acorn User
Show. He told me that quite definitely I should use the Epson LQ860. So
this I do, and without any problems so far. I haven’t even edited out
the control codes for the colours − the driver is used, unmodified. (Ken
Cowap)
6.3
Where are the hare and the tortoise?
6.3
Can I get the hare and tortoise icons that came with my ARM3 back on the
iconbar? Or do I have to tell people that if they want E-Type to go at a
playable pace they have to press <F12> then type “Cache off” then press
<return> twice before running it?! (I just use
6.3
RMEnsure Arm3Icon 0.9 Run scsi::4 .$.<leafname>.!Arm3
6.3
in my !boot file and... hare presto! Ed.)
6.3
Programmers’ Reference Manual
6.3
At the Acorn User Show, a “beta” copy of the new PRM was on sale in
photocopy form for £120. The final version will apparently be available
“in the first quarter” of 1993. It is about 50% bigger than the RISC-OS
2 manual and will fill six volumes. From my brief inspection, it looks
as if it contains quite a lot more explanatory material than its
predecessor. I was told that neither the price of the final version nor
details of the trade-in arrangements (if any!) have yet been fixed.
6.3
Various hints and tips:
6.3
• Problem with Impression dongle − You can’t write to a printer when the
Impression dongle is fitted unless you have run the Impression appli
cation first. (Mike Williams)
6.3
• Summer time blues − Some C programs (e.g. PVray) get the time wildly
wrong when *BST is configured. (Mike Williams)
6.3
• Bug in !SciCalc − It gives wildly inaccurate results to the <±>
function for some values, e.g. if you enter “1<Exp>56” then press <1/x>
to give “10E−57” then press <±> you get the surprising result: “10E−58”.
If you press <±> again, you get “10E−41”. (Mike Williams)
6.3
• Where did this file come from? − There are quite a few ways in which
<adjust> clicking can be very helpful. I’m sure you all know that
<adjust> double-clicking a directory opens it and closes the parent and
that using <adjust> to close the directory opens up the parent. You may
even know that clicking the close icon with <adjust> while holding down
<shift> actually opens the parent without closing the directory. What
you may not have noticed is that you can do the same with (some)
applications. With most applications, it seems, <adjust> double-clicking
a file loads the file and closes the directory from which it came. (I
think this is a function of the Filer and is therefore true for all
applications and files. HE.) Then, if you click in the close box with
<adjust>, holding <shift> at the same time, the directory from which the
file came appears on screen without the file disappearing from view.
With Impression, however, there is no need to hold <shift> as well −
just using <adjust> on the close box brings up the directory viewer in
which the file is saved.
6.3
Oh, and I’ve only just discovered that if you draw a window’s scroll bar
with <adjust>, you can move about in both x and y directions! (Paul
Beverley)
6.3
• Rubber-banding − Those who have gone straight to RISC-OS 3 without
reading the manual (like Ed!) may not have noticed that you can use
rubber-banding to select a group of files. So instead of <adjust>
clicking every individual file you want to select, you can click
<select> in the space between two file icons and, as you drag, a red
bounding box appears which you can release when it is over the files you
want. Using <adjust> instead of <select> simply adds files to those
already selected but note that if you go over a file that is already
selected, it is then de-selected. (Paul Beverley)
6.3
• SerialDeviceDriver module − On an A5000, never unplug the module
called “SerialDevice Driver”, as the following strange things will
happen. After a power off/on the machine will boot up but, during
booting, an error message “Error from Task Manager” will appear and
there is only a “Cancel” button to click on. Then the boot up is
finished correctly but no Taskmanager is present! It will only reappear
if you double-click on any file with the new Desktop filetype (&FEA),
even if this is 0 bytes long! In addition, the Taskmanager will not
function correctly in all cases: try to get an “!Edit Task Window” now.
Bug or not? (Carsten Bussman)
6.3
• Double-clicking into Basic − On my first machine (which was killed by
my dealer while trying to fit a SCSI hard drive) I was able to enter BBC
Basic V directly by double-clicking into the system’s start-up message
window (“RISC-OS ... initialising”). This does not work with my current
machine (Serial no. 1009939); I don’t know why. This was a nice feature
which I’m missing now, so why isn’t it implemented in all RISC-OS 3
ROMs? (Carsten Bussman)
6.3
• Inserting path names into !Edit − Try dragging any file or directory
into an !Edit window while pressing the Shift key. This will give the
file’s pathname. (Carsten Bussman)
6.3
• “Menu Area Full” error in FormEd – Captain K. J. Tompkins reports the
error “Menu Area Full (Internal code 6560,1)” from his version of
!FormEd. This is probably because he has one of the earlier versions of
this program which only allows limited space for building its menus and
is defeated if there are too many fonts installed. If so, it is fairly
easy to cure: in the Basic program called !RunImage inside the !FormEd
directory there is a line that reads something like:
6.3
DIM menufree% &1800: menuend% = menufree%+&1800
6.3
(this is at line 790 in the version I’m now looking at). If each of the
two &1800’s is replaced by &2000, so that the line reads:
6.3
DIM menufree% &2000: menuend% = menufree%+&2000
6.3
the program should work.
6.3
(Note: this hint does not apply to all versions of FormEd) (Hugh Eagle)
6.3
• Filer_Run syntax − it seems that the command
6.3
*Filer_Run <App$Dir>
6.3
(where App$Dir is a system variable set to the name of an application)
doesn’t run the application as you might expect. What it does is to open
a filer window on the application’s directory with “<App$Dir>” (not the
actual pathname) in the title bar.
6.3
The following syntax, however, does seem to work:
6.3
*Set Alias$FR Filer_Run <App$Dir>
6.3
*FR
6.3
*UnSet Alias$FR
6.3
(this is equivalent to the old trick that used to be needed in RISC-OS 2
to open a directory viewer with *Filer_OpenDir). (Hugh Eagle)
6.3
• *FontRemove syntax − The FontRemove star command seems to be very
fussy. You have to tell it the name of the directory that you want to
remove in exactly the form in which it is included in the Font$Path
system variable. For instance, if you want to remove
RAM::RamDisc0.$.!Fonts. from Font$Path you must type
6.3
*FontRemove RAM::RamDisc0.$. !Fonts.
6.3
including both the disc name (RamDisc0) and the full stop at the end.
(Robert Chrismas has sent in a handy application called !FontsCut which
lists the directories currently included in Font$Path with numbers
beside them and all you have to do is type in the number corresponding
to the directory you want to remove. This is included on the monthly
program disc.) (Hugh Eagle)
6.3
• Configuring Shift Caps − In the Caps lock section of the Keyboard
window of !Configure, the effects of “Caps lock on” and “Shift caps”
seem to be reversed. In other words, selecting “Caps lock on” means that
holding down <shift> reverts to lower case whereas selecting “Shift
caps” means that you get upper case regardless of whether <shift> is
held down or not. (Hugh Eagle)
6.3
• Mouse button codes – If you enter and run the following simple Basic
program:
6.3
REPEAT
6.3
MOUSE X,Y,B
6.3
PRINT TAB(10,10)B
6.3
UNTIL FALSE
6.3
then watch what number is displayed on the screen as you press the
various mouse buttons you will see:
6.3
no button 0
6.3
Select 4
6.3
Menu 2 for an instant, followed by 0
6.3
Adjust 1
6.3
The behaviour of the menu button has changed from RISC-OS 2, I think.
Those programs that are interested only in detecting whether the button
has been clicked won’t be affected, but the change can cause problems
for programs that want to detect whether the Menu button remains held
down. Luckily, there is a simple solution: the function
6.3
INKEY(-11)
6.3
will return the value TRUE however long the Menu button is held. (Hugh
Eagle)
6.3
• Getting Art Nouveau to run – Hands up those who remember Art Nouveau?
Well, this excellent old program has one quirk, namely that you have to
hold down the menu button while you move round the menu tree until you
make your selection. So, because of the changed functioning of the mouse
button (see above), it doesn’t work in RISC-OS 3.1 unless you make a
slight modification. In the Basic program called ArtNouveau from the
directory ANfiles at line 30950 in my version there is a line that
reads:
6.3
WHILE stayopen AND mbut=2
6.3
I have found that altering this to:
6.3
WHILE stayopen AND INKEY-11
6.3
restores the program to health. (Hugh Eagle)
6.3
• Using a TaskObey file to display a message – The new filetype
“TaskObey” is useful if you want the !Run file of an application to pop
up a message in a window without halting the whole system. You might use
this for instance to display a warning to change the setting of your
printer. Thus, to display “This is a message” in a window, create a file
called, say, Message containing:
6.3
Echo This is a message
6.3
set the filetype to TaskObey and include, in the application’s !Run
file, the line
6.3
Filer_Run <Obey$Dir>.Message
6.3
If you had set the filetype to Obey, the message would pop up moment
arily then disappear before the application loaded. (Hugh Eagle)
6.3
• Deskboot files and the Pinboard – The User Guide explains how to
create a !Boot file to boot the computer in your desired configuration.
It adds that, when saving the !Boot file, information to start up
applications produced before RISC-OS 3 is not saved automatically. To
get round this, it explains how to rename the !Boot file as Deskboot and
to incorporate it in a !Boot application. It goes on to explain how to
edit the Deskboot file, either to remove or to add lines as necessary.
6.3
If you use the Pinboard as I do, some icons will be changed on a regular
basis. For example, I have regularly used files which change each month.
If the Deskboot file is saved each time the Pinboard is changed, a
certain amount of what the User Guide calls “hand editing” would have to
be done each time a change is made, to remove or add relevant
instructions.
6.3
There is an easy solution. Set up your Desktop once and save the
Deskboot file. Then click <menu> over an empty space on the desktop and
save the Pinboard settings in the Pinboard file in your root directory.
Now drag your Deskboot file into !Edit (or Shift-double-click on it) and
delete all the lines which begin “Pin”. Alter the line “Pinboard” or
“Pinboard -Grid” to read “Run Pinboard” or “Run Pinboard -Grid”.
6.3
Future changes to the Pinboard setting now need be saved only to the
Pinboard file in your root directory. (John Waddell)
6.3
Ken Cowap warns that a DeskBoot file can easily be unnecessarily
cluttered if, for instance, there are a number of unwanted items on the
screen when it is made. This will incur a penalty in the time taken to
boot up − his started at two and a half minutes before he trimmed it
ruthlessly.
6.3
• ADFS buffers − According to Acorn’s telephone help line, the bug which
made it advisable to configure the adfs buffers to zero has been fixed!
(According to Computer Concepts, unless the buffers are configured to
zero, it is vital to dismount disks before removing them for good from
the machine, because otherwise there may be trouble later.) (Jochen
Konietzko)
6.3
• OS_Byte calls and the serial port – I use my A3000 for many purposes,
one of the more important being as a terminal to our mainframe computer.
The software for this I have written myself in Basic. After the upgrade,
my terminal program wouldn’t work. It would not accept characters from
the serial port. I tried using the serial port utility module supplied
with the upgrade, but this did not help.
6.3
After a lot of experimenting, we finally discovered that if we changed
all the *FX calls for SYS “OS_Byte” calls, then the program would work.
It still puzzles me why two supposedly similar commands should give
different results! (S. J. Furnell)
6.3
• Z88 Filers – The Release Note says that the !Z88 filer supplied with
Pipedream requires interactive filing to be off. You’ll be pleased to
know that David Holden’s !Z88 filer (Careware 7) suffers from no such
problems. (Steve Hutchinson)
6.3
• Programs that call machine code from Basic – An application I used a
lot, !BPrint (Risc User) wouldn’t run with RISC-OS 3, and the cure for
this may well apply to other applications organised in a similar layout,
i.e. using a central Basic program calling a separate piece of machine
code. The cure was to re-assemble the code under RISC-OS 3; obviously
there is some small difference between its assembler and that of RISC-OS
2. In this particular application, its “ancestor” (the program developed
over a period of time), a program called FileList from vol.2 issue 5 was
rerun to produce the assembled file BLIST which was substituted for the
original. (Ken Cowap)
6.3
• Mode number conflicts – A program for plotting Mandelbrot diagrams
failed with RISC-OS 3 and the manner of its failure may indicate a cure
for others. This particular program called a Mode 40 which didn’t exist
under RISC-OS 2 but which was specially designed and provided with the
program. RISC-OS 3 of course does have a native Mode 40 and a clash
between it and the “special” gave rise to the error report. The cure was
to rename the “special” to a mode number outside the range provided with
RISC-OS 3, in this case Mode 50. (Ken Cowap)
6.3
• Setting !Alarm options – This is a minor matter, unless you are trying
to set your own !Alarm display format in a !Boot file. This you do with
a Set Alarm$Options line (p136 App Guide). The quotes are omitted. This
is my (working) line:
6.3
Set Alarm$Options -format “%z24:%mi %zdy.%zmn.%zyr”
6.3
If you use a !Boot file and !Alarm, put it in to see my favourite
display format! (Steve Hutchinson)
6.3
Miscellaneous comments
6.3
• Surely a ‘grown up’ operating system shouldn’t say, “Disc error 10 at
:0/00000000” when you try to read an unformatted disc?
6.3
• It seems strange that !Configure requires you to press <return> after
typing into a writable icon. This is not normal RISC-OS convention.
6.3
• It’s a shame that they didn’t manage to squeeze some suitable backdrop
sprites onto one of the RISC-OS 3 discs. Those who have access to
Windows 3 might like to know that its wallpaper bitmaps seem to work
very well on the RISC-OS 3 Pinboard, and it is very easy to transfer
them from an MS-DOS floppy, via !ChangeFSI (on the Support disc), into
sprite format on an Archimedes disc. The file called “leaves.bmp” makes
an excellent seasonal backdrop (I’m writing this in October!).
6.3
Alt-key combinations
6.3
The last couple of issues of Archive have contained lists of Alt-Key
combinations in RISC-OS 3.1. To complete the picture I would add:
6.3
To get accents ( ´ ` ¨ ^ ° ¸ ) on their own, press the appropriate Alt-
key combination (e.g. Alt-[ for a ´ ) then the space bar.
6.3
Alt-Spacebar gives a “hard space” or “non-breaking space” (the sort you
put between “NR2” and “2SD” if you don’t want Archive’s postcode
“NR2 2SD” to be split over two lines.)
6.3
I have put a diagram of the keyboard showing the Alt-key combinations
below, and I hope that Paul will put the drawfile on the monthly program
disc.
6.3
Incidentally, I am a bit puzzled about the character (code 173) that is
produced by the Alt-hyphen combination. What’s it for? I thought at
first that it was the “soft hyphen” that is described on page 109 of the
Impression 2 manual, but I now see that this is produced by Ctrl-hyphen
(this character – code 31 – is normally invisible but indicates a
preferred place for breaking a word with a hyphen at the end of a line;
it works even if the hyphenation module is not present). By contrast,
character 173 is, so far as I can see, always visible (as a hyphen if
the font follows the Acorn standard) but its effect is to prevent a word
into which it is inserted ever being broken over the end of a line even
if hyphenation is switched on. Am I correct in describing it as a “hard
hyphen”? (Well, I can think of one application. Suppose you want to use
a word like “prayers” where you deliberately hyphenate the word to
refer to the people doing the praying. In that case, if it were an
ordinary hyphen, the word might get split at the end of a line. Using a
‘hard hyphen’ avoids the split. Ed)
6.3
The upgrade again
6.3
Finally, I thought it would be good to take a slightly light-hearted
look at the subject as we hear the experiences that Rob Wears had in
fitting and using RISC-OS 3 on an A310...
6.3
I finally managed to get my grubby mitts on the RISC-OS 3 upgrade at the
Acorn User show, and arrived back home with that odd mixture of terror
and excitement that comes from having to take the lid off the machine.
I’ve fitted one or two things in the past, like the soldered 4MByte RAM
upgrade, but no matter how many times I do it, it’s still a bit like
performing neurosurgery on a close friend. Will Archie pull through,
Doc? Actually, fitting was a lot easier than I’d thought. I used the
IFEL carrier board, as their memory upgrade has proven so reliable. My
only source of concern was the discrepancy between their fitting
instructions and the ones supplied by Acorn which alluded to changing
links on the mother board. I followed the IFEL guidelines, which had the
link changes on their board, and waited anxiously for the machine to
explode when I switched it on. No problem, apart from the fact that I
bypassed the desktop and ended up at the command line. Still, the
desktop was easily accessible with a simple “*” command.
6.3
The hard drive icon was where it should be (and no longer jumped around
when I mounted or dismounted the disc!) and I am pleased to say that I
encountered no problems with the Oak SCSI board (version 1.16 of SCSI
driver).
6.3
I couldn’t resist the urge to delve into the Apps directory, and here I
encountered my first problem. Initially, I couldn’t get !Edit to load
from Apps and was greeted with the wonderfully informative message ‘File
“Basic” not found’. Fighting the sudden tide of panic (“It wouldn’t work
at all if the ROM legs were snapped off, idiot!”) I checked the module
list via the command line and found that I was apparently the lucky
owner of the one set of chips which Acorn had shipped without putting
Basic on board first. I was about to fall on my soldering iron in
despair when I thought of checking the status of the system modules
using *ROMModules and found that the Basic module was unplugged. I
reinitialised the module, and !Edit worked fine. There, that wasn’t too
bad − the hair will probably grow back and nobody will notice the
teethmarks on the manual....
6.3
Incidentally, the ARM3Support module was also unplugged and therefore
the *Cache command wouldn’t work, but I’d got the hang of things by now,
and pressed on happily. I then UNPLUGged the modules that I thought I
wouldn’t need (old habits die hard) like the Econet and NetFS modules −
each unplug command was greeted with the comment “xxxx module not found”
yet a subsequent check revealed that the module had been unplugged
correctly − odd! Archie hadn’t behaved like this before his lobotomy. It
was a bit unnerving, like suddenly finding the Queen Mother skateboard
ing down your front drive.
6.3
I then set about gingerly reconstructing the !Boot option for the hard
disc, feeding the bits to my new Archie in small, digestible chunks.
Compression (version 1.10) and !Spark both worked fine, after I
remembered not to set the “Newer” option for copying. However, Archie no
longer likes the “OS_UpdateMEMC” call which I had used to speed up the
ROMs. This is now treated with the same degree of horror as if I’d just
offered it deep-fried weevil for tea, and the machine has to be kicked
out of its sulk with a hard reset. I initially and unjustly blamed the
old OSSys module, but this actually seems to be fine and is the only
real survivor from my old collection of useful bits and pieces − have
they STILL forgotten to put a *SYS command in?
6.3
The next little gem was the “ChangeDynamic Area” calls which are
automatically saved in the desktop !Boot file − these override the CMOS
RAM settings and I couldn’t work out why the system sprite area remained
so large despite all my efforts to shrink it. And remember, Gentle
Reader, not to save the !Boot file in a Compression directory because,
despite what they all say, those teethmarks in the manual do show! On
the brighter side, I found that the configured startup language was not
the desktop, which explained Archie’s new enthusiasm for the command
line. The number for the configured language of your choice is now
obtained by counting down the list of modules − how simple. Not a
mention of alchemy or the Black Arts. Good grief, if it’s that simple,
anybody will be able to do it...
6.3
The !FontsPlus application caused some frontal headaches (actually,
these may have been caused by sudden and repeated violent contact
between the forehead and the monitor) as I could not persuade it to
recognise the ROM based fonts. Off it went to the great filestore in the
sky, and back I went to using multiple directories with original names
(!Fonts, !Fonts2, !Fonts3, !Fonts4, etc.). How quaint!
6.3
On to some serious computing. Impression works! (Actually, I cheated. I
got the upgrade from those nice CC people at the show). And there are
more screen modes available from startup, and a *WimpMode command to
simplify my DTP !Boot sequence. On the down side, I have found printing
to be a bit of a pain − my old !PBuffer application doesn’t work anymore
and even making the PrinterBufferSize HUGE using *CONFIGURE doesn’t seem
to do anything. However, there are many more options available for
printed output. I can even get 360×360 dpi on the Panasonic and it only
takes about 30 minutes a page!
6.3
I seem to have more space on the hard disc now. I used to have lots of
little applications that I’d collected over the years which now have
their functions in ROM, such as SetType, FileFind, StickyBD and
ShutDown. Being ruthless, and having no more need for them, I erased
them completely. Who knows, one day, I might even erase the backups of
the backups.
6.3
So am I sorry to have upgraded? Not at all. It was just that the scale
of the change surprised me. I suppose that the change was more compli
cated because I had my system set up exactly the way I wanted it, and
had to restructure virtually the whole hard disc. It was just as
traumatic as upgrading from Arthur to RISC-OS2 had been, but I’m very
pleased with my new machine. Of course, 3.1 goes into 310, and it may
even be 100 times better!
6.3
(Actually, this has an important moral − if Rob had followed Acorn’s
instructions and done a <delete-power-up> in the first place, he
wouldn’t have had half the problems he did. Ed.)
6.3
And finally ...
6.3
Send your comments, suggestions, queries, complaints, hints and tips
either to me, Hugh Eagle, at 48 Smithbarn, Horsham, Sussex RH13 6DX. A
6.3
The characters shown in bold are produced by pressing the relevant key
in conjunction with either the Alt key on its own or both the Shift and
Alt keys:-
6.3
e.g. to type ¢ hold down Alt and press the C key
6.3
to type © hold down Shift and Alt and press the C key.
6.3
The characters indicated by an arrow are intended to be used in
conjunction with other keys to produce accented letters; in each case
the accent is only produced if the Alt-key combination is followed by an
appropriate letter: e.g. to produce é you first type Alt-[ then type e.
If you want the accent on its own, follow the Alt-key combination by
pressing the spacebar.
6.3
Alt-<hyphen> produces a “hard hyphen” (character 173)
6.3
Alt-<spacebar> produces a “hard space” (character 160)
6.3
6.3
Small Ads
6.3
(Small ads for Archimedes and related products are free for subscribers
but we reserve the right to publish all, part or none of the material
you send, as we think fit. i.e. some people don’t know what ‘small’
means and there are certain things, as you can imagine, that we would
not be prepared to advertise as a matter of principle. Sending small ads
(especially long ones!) on disc is helpful but not essential. Ed)
6.3
• A3000 external drive buffer (Beebug) £20. ArcDFS £12. Phone
0502−565568.
6.3
• A3000 RISC-OS 2, 4Mb RAM, serial port, Acorn DTP, Phillips CM8833
monitor & stand. £650. Phone 0235−813557.
6.3
• A3000 2Mb RAM, socketed ARM2, RISC-OS 3.1, VIDC enhancer £500, 45Mb
Hardcard £299, PRES A3K6 second drive i/f £20, PRES A3K4 double exp. box
£39, Learning Curve s/w (proceeds to charity) £100. Phone Barry on
0332−701969.
6.3
• A310M, colour monitor, PipeDream, Graphic Writer, etc, £500. Gordon
Walbridge on 0305−782623.
6.3
• Acorn Desktop C comprising Acorn Desktop Development Environment and
ANSI C release 4. As new and unregistered £160. 0272−736237.
6.3
• Amstrad FX 9600T fax machine £175. Tandy FP215 flat bed plotter £149.
Integrex Colourjet 132 printer with new cartridges and paper £299. Phone
Stuart on 0792−204519.
6.3
• Beebug 5¼“ disc interface, 40 track drive, DFS reader software and 50
5¼” discs. The lot for only £50 + p&p. Phone 0332−557751.
6.3
• CC ROM/RAM podule with 6 × 32Kb RAM and 1 × 8Kb RAM and battery-
backup. £50 o.n.o. Phone Thomas on 45+31−67−89−00 (Denmark).
6.3
• CC Turbo driver for BJ10ex never used (unwanted present) £35. Phone
Chris on 081−856−4732.
6.3
• Digitising tablet TDS LC12 professional quality tablet with a 12“⇧×12”
active area, 4 button puck, p.s.u. and software to use instead of a
mouse in Draw, etc. £110 o.n.o. Phone 0903−265727 eves.
6.3
• Flexifile £40, Acorn DTP £35, First Word Plus 2 £30, Inertia £5, E
Type £5, Bug Hunter/Moon Dash £5, Arcpinball £10, Mad Prof. Mariarti £8,
Pipemania £8, Nevryon £8, Elite Gold £20, The Last Ninja £15, Star/
Epson/Citizen Colour Printer Driver £10, Panasonic KX-P1081 Printer £40.
Phone 0277−625613.
6.3
• FontFX £2, DrawBender £2, Glimpse £3, Help Companion £3, Mad Prof.
Mariarti, Pacmania, Arcade Soccer, Fish, Fireball II £5 each, FunSchool
2 (u6) £6, (6-8) £6, Droom, Personal Accounts, Interdictor 2, Apocalypse
2 £10 each, PenDown £25. Phone Dave on 081−292−6870 (eves).
6.3
• Games. Chopper Force £15, Star Trader £3, Pacmania £3, Worra Battle
£3, Terramex £3. Phone Chris on 0276−20575 after 6pm.
6.3
• Laser Direct Qume + 2 spare toner cartridges, vgc + latest software,
£600. Phone Ray on 081−864−7208.
6.3
• Morley 1Mb RAM (upgradable) for A3000, £40. PRES podule case £5. Acorn
ADF32 medium res colour monitor, £50 + carriage. Micro User 1.1 to 10.2
(complete), most in binders, £10 + carriage. Input (complete) in
binders, £10 + carriage. Phone Pete on 0603−897511 or CompuServe
70374,207.
6.3
• Multipod podule for Archimedes A300/400 series. Includes serial port,
video and sound digitiser/analog port and 3 ROM sockets all for £60.
Phone Clive on 0908−661980.
6.3
• PC emulator V1.8 £50, Poster V1.4 £25, Lincad Drawing program V1.45
£25, Beebug Hard Disc Companion V2.0 £20, Beebug colour printer driver
£7. Phone 0536−724981.
6.3
• Pineapple colour digitiser hardly used, accept £150 (cost £232). Also
Leading Edge joystick interface (A5000 compatible) plus 2 Superpro
joysticks, unused £45 (cost £62). Phone 0278−751317.
6.3
• Risc Developments memory board (4Mb) for A310. Will take RISC-OS 3.1.
Unused £95. Phone 0423−565997.
6.3
• Scanner − Watford Mk I hand scanner, up to 400 dpi, £60. Beebug 5¼“
disc interface Mk II, £20. Phone Alex on 0705−526800.
6.3
• Schema V1.18, boxed with manuals. £50 o.n.o. Phone Mr I Mackay on
081−969−7294.
6.3
• Shinwa CPA80 printer £50; Impression Business Supplement £20; Games −
Elite, Swiv, Zarch, etc. Phone Rob Wears on 021−384−4152.
6.3
• Sony 3000p video camera. Ideal for video grabbing! Powered from VCR or
simple 12v supply. Only £100 o.n.o. Contact Bob Ames 071−477−8275 day or
0487−814227 eves, w/e.
6.3
• Wanted 4 to 8Mb upgrade for A410/1. Interested in almost anything,
even just a circuit diagram that I can build myself. Phone Thomas on
45+31−67−89−00 (Denmark).
6.3
• Wanted, ARM3 processor and Impression II. Phone 0332−557751.
6.3
• Z88 in case with custom rechargeable system, extra RAM, PC/Acorn link
EPROM, Parallel printer cable, Comms link EPROM, magazines £175. Phone
0452−618742.
6.3
Charity Sales − The following items are available for sale in aid of
charity. PLEASE do not just send money − ring us on 0603−766592 to check
if the items are still available. Thank you.
6.3
(If you have unwanted software or hardware for Archimedes computers you
could donate for charity, please send it in to the Archive office. If
you have larger items where post would be expensive, just send us
details of the item(s) and how the purchaser can get hold of them.)
6.3
A3000 1Mb upgrade £30, First Word Plus 2 + Study Guide £30, Diet Manager
£5, A4 Forms Designer £5, Hearsay £15, Pace Nightingale modem (300/300,
1200/75) £25, Artisan Support Disc £5, ArcTiculate £15, !Help − Intro to
Archimedes £5, Snippet £10. A
6.3
Oak Solutions
6.3
From 6.2 page 6
6.3
Music Column
6.3
Stewart Watson
6.3
The range of music software available for the Archimedes has been
expanding at a fairly acceptable rate over the last couple of years.
Although there is unlikely ever to be the same volume of packages
available as there is for the Atari ST or PC compatibles, we should take
comfort from the fact that the general standard of software for the
Archimedes is consistently high.
6.3
Music packages fall into three general categories:
6.3
1. Packages using the internal sound chip for backing tracks and
sound effects for programs
6.3
2. Sequencing software, usually using MIDI keyboards and expanders
6.3
3. Music scoring software
6.3
Some programs overlap two, or even all three, categories but, when
buying software, it is essential that you keep in mind its primary use.
For example, although a notation package might have sequencing facili
ties, it is unlikely to have all the facilities found in a dedicated
sequencer. Below I’ve listed some of the major packages and commented on
their current states of development.
6.3
Shareware
6.3
There are now stacks of music files available from the various shareware
suppliers. Most of these are Tracker files, many of which are ported
across from the Amiga. Tracker players, of which there are several in
the public domain, are all that is required if you want to play Tracker
files. (Tracker itself is required if you want to write files.) Cream PD
specialises in supplying Tracker files − £1.00 for a catalogue which
contains some useful Tracker utilities, including a Tracker player, and
some nice demos. Tracker itself can be obtained through Archive for £47.
6.3
MIDI files
6.3
At the other end of the spectrum from shareware tracker files are Hands-
On Midi Software, who supply chart hits and classical pieces in MIDI
file format at about £6 for a chart hit or £12 for a classical piece.
These pieces are professionally sequenced and highly recommended. A
demonstration disk is available for £5.95. If you order any of these, be
sure to specify Archimedes and Midi file type 0 or 1.
6.3
Electro-Music Research
6.3
Mike Beecher’s company, EMR Ltd (Elecro-Music Research), was one of the
first to produce music software for the Archimedes. Mike is still going
strong and waxing as eloquent as ever. Studio24+, a sequencing program,
is now up to Version 3 and has a massive number of new features
including improved scoring, CD-ROM control, multiple reels, MIDI
instrument list, reverberation control, etc. I’ll look at Studio 24
Version 3 in depth in a future article.
6.3
Sound Engineer, which should be released by the time you read this, is a
totally rewritten and revamped version of SoundSynth − a program used
with a sampler podule or the internal sound chip to create sounds.
6.3
EMR have a whole suite of music packages for the Archimedes and further
information and a catalogue of all the available programs can be
obtained from Electro-Music Research.
6.3
Clares
6.3
Clares came later onto the scene but have made rapid progress. They now
have Rhapsody II, a score writing package; Armadeus, a sampler;
ScoreDraw which converts Rhapsody files into drawfiles; Vox Box, another
sound creation program; and the newly launched Rhythm Bed, a drum
sequencer. These are all available from Clares Micro Supplies.
6.3
I believe that they are about to announce the launch of a real-time
sequencer.
6.3
Scorewriter
6.3
The score writing package with the most flexibility is undoubtedly
Scorewriter. Unfortunately, the price of flexibility is user-unfriendli
ness. I vividly remember the first weekend I spent with Scorewriter − I
almost threw the computer out the window! However, once you have
mastered ScoreWriter, it is an absolute joy to use.
6.3
Scorewriter can be obtained from EMR or the author, Philip Hazel 33
Metcalfe Road, Cambridge CB4 2DB. There is a completely new RISC-OS
version which will work from the desktop, available around Christmas. It
is tremendous, but not for the faint-hearted.
6.3
Longman-Logotron
6.3
Longman-Logotron, well known for their involvement in Education, have
just released Notate Version 1.2a. This has added real-time record
facilities to Notate, together with some other extra features. One of
the big pluses of Notate is the number of template files provided with
the program to save the user time in setting up.
6.3
Inspiration
6.3
Inspiration is a 250 track high specification sequencer, originally by
Pandora Technology but now marketed by Amp Sound. It has attracted a
fair following, and its files can be transferred to Rhapsody for
printing. Unfortunately, it comes with a dreaded dongle.
6.3
Summary
6.3
As you see, the amount of software available is steadily growing and the
quality is improving. As long as Acorn pursue a sensible pricing policy
and manufacturers follow a sensible protection policy (which in my book
does not include dongles), the market for both hardware and software
should continue to grow apace. A
6.3
Help!!!!
6.3
• A4 concept keyboard wanted − Has anyone got an old A4-sized concept
keyboard (for a BBC B computer − not an Archimedes) they don’t want?
It’s for a downs syndrome lad whose parents work for a Christian charity
and are therefore not too flush for cash? If so, please ring Paul
Beverley at the Archive office or Dave Carlos (a friend of their family)
on 021−777−8957.
6.3
• Amiga formats − Does anybody know of an application that allows the
use of Amiga-formatted discs with an Archimedes? Contact Graham Crowe at
Droitwich High School, Ombersley Road, Droitwich WR9 0AA.
6.3
• Draw to Paint − How can I convert Draw files into sprites? Roger
Darlington, Manchester.
6.3
• Impression − Some DTP packages don’t actually include graphics files
within the document but simply contain a pointer (filename) to where the
graphic is being stored. Impression does not work in this fashion; every
component of the document is contained in the document’s ‘application
directory’. In general, I am sure this is the most sensible thing to do.
However I have a problem. I keep the letters I write at work on disc and
I am running out of space. Each letter contains a copy of the College
crest, whose drawfile occupies about 10 Kb, or 5 Kb if compressed − this
is often much larger than the text content of the letter! Is there any
way for the Impression document directory to hold just a reference or
pointer to the crest? I know that a solution would be to store just the
text story of the documents, but that would be rather inconvenient. Any
other solutions? Brian Cowan.
6.3
• JPEG − Does anyone know how to get “JPEG”, a PD JPEG image conversion
program, to work under RISC-OS 3.1? Or is there another PD JPEG program?
Roger Darlington, Manchester.
6.3
• Missing charity discs! − At the Acorn User Show, Martyn Diplock bought
a second hand copy of System Delta Plus version 2. Unfortunately, there
were no discs in the box − just the documentation! If you are the person
who donated the software and still have the discs, could you possibly
send them to Martyn at 162 Willingdon Road, Eastbourne, East Sussex?
(Martyn said that the £20 he paid could stay in the charity pot − which
is very generous of him − but it would be nice to get him a working
version of SDPlus 2.) Thanks. Ed.
6.3
• Serial port speed − Does anyone know of a way of getting a serial port
working on an A410/1 at more than 19,200 baud? Philip Lardner, Co.
Wicklow, Ireland.
6.3
• Stolen computers − Watch out for an A420 serial number 1004092 and an
A3020 serial number 1010239 − they were stolen from Bromsgrove School!
If you see them, please inform the police a.s.a.p.
6.3
• TableMate − We sent someone a copy of TableMate for review but forgot
to record who it was! Would that person own up as we have a new up-dated
version to send you. Thanks. Ed.
6.3
• Wordwise Plus − Is there anyone in the Milton Keynes area who could
help a disabled lady with getting Wordwise Plus programs working on her
BBC Master? We asked for help some time ago and a gentleman responded
and was very helpful. Unfortunately, he has moved to London! Contact
Marjorie on 0908−607012. (Please allow time for her to get to the phone
and answer it as she is not too mobile.) Thanks. Ed. A
6.3
PD Column
6.3
David Holden
6.3
Anyone who is interested in Archimedes PD will probably have heard of
Data Stream. What you may not know is that it has recently moved. It’s
now run by Phil Coleman and the new address is 34 Norbreck Close, Great
Sankey, Warrington, Cheshire WA5 2SX. If you want to know more then send
£1 for a catalogue.
6.3
PD prices
6.3
Many PD libraries are having a hard time at the moment. Costs are
rising. For example, thanks to the recent devaluation I now have to pay
an extra £6 per hundred for discs. Since libraries are ‘small busi
nesses’, like others in that class we are expected to help the high
street banks recover from the consequences of their bad decisions by
paying high charges. Most people who have only personal bank accounts
probably don’t realise that when you have a business account you not
only pay a quarterly service charge but also 60−70p per transaction.
This is one of the reasons why libraries normally give credit notes
instead of refund cheques. A 50p refund cheque could cost the library
about £1.20 by the time the bank charges are included.
6.3
The newer libraries have been cushioned from this because most banks
offer free banking for the first year for a new business, but this time
has now run out for those which started up last year and they are
beginning to feel the pinch.
6.3
One way of bypassing bank charges for small orders is to avoid cheques.
The problem is that, unlike the old £1 note, coins tend to be rather
obvious and also heavy. Postal orders are not only inconvenient for the
sender, who must go to the Post Office to purchase them, they also
involve a premium.
6.3
There is one item which is small, light and is as good as cash − stamps.
Most Archimedes PD libraries are not large enough to employ a Post
Office franking machine, so we all use vast quantities of first and
second class stamps. All the libraries that I have spoken to would be
happy to accept, for example, £1 in stamps instead of a £1 cheque for a
catalogue.
6.3
As a result of these general increases in costs I expect that some
libraries will soon be increasing their prices. Many have been trying to
operate selling discs at £1 each. If you don’t allow anything for
overheads or your own time, it is just about possible to cover net costs
at this price. Remember that sending a disc doesn’t just involve the
cost of a disc, a stamp and a disc mailer. There is also the wear on
your computer, printer consumables, the occasional dud disc, letters and
free discs to Authors, and so on, which all form part of the basic
overheads of running a library. And that’s before we get down to less
obvious things like paying for virus detection software, extra hardware,
and, of course, bank charges.
6.3
Prices down
6.3
It’s not all bad news on the price front. As announced last month in
Archive, the price of the ‘Shareware’ PD discs has been reduced to £2
and the Careware to £5. This has prompted me to reduce the price of
A.P.D.L. discs as well. From now on, they will be priced at £1.50 each,
with one free disc of your choice for every five purchased in place of
the previous percentage discount scheme.
6.3
Maastricht
6.3
I hope I still have your attention. Just that one word may be enough to
make you skip the rest of this article.
6.3
I’m not going to get political but until now it’s been almost impossible
to find out exactly what this is all about. Everyone is talking about it
and almost no one, including the politicians, has actually read it. I
have a copy on disc and so if you want to discover what it really says
then send £1 (or four 1st class stamps, of course), to me at the address
below. Don’t just listen to what the politicians tell you it means, read
it for yourself and make up your own mind.
6.3
Disc compression
6.3
Most consumers of Archimedes PD will be aware of ArcFS used by almost
all PD libraries. What you may not be aware of is that there is a
similar Shareware program available for the PC. This is invaluable for
PC Emulator users because unlike the expensive commercial options,
Stacker and SuperStor, it works fine on floppies and you can mix
compressed and uncompressed files on the same drive and directory. The
program is called SLIM and since it’s not American, most PC Shareware
libraries have never heard of it.
6.3
As with ArcFS, you need to compress the files first and then the
resident portion of the program, normally loaded in your AUTOEXEC.BAT
file, decompresses them ‘on the fly’ as you load them. The decompression
is so fast that you probably won’t even notice it is happening using the
Emulator. This often makes it possible to run programs from a single
floppy disc that would otherwise need either a hard disc or lots of
swapping.
6.3
Ray tracing again
6.3
Another new program has appeared. This is called DBW Render and looks
like the best yet. It was ported from the Amiga and the version that I
have also contains the ‘C’ source code.
6.3
I’m still hoping that someone will offer to do a comparative ‘in depth’
review of the various PD packages because there are now several
available and they all have different good and bad points.
6.3
Authors’ association?
6.3
Possibly as a result of the things I have been writing recently in this
column, there seems to be an awakening interest in the concept of
Shareware amongst PD authors. I have received several letters asking if
there is any sort of association for Archimedes PD/Shareware authors
and, if not, why not?
6.3
It certainly seems like a good idea. The Association of Shareware
Professionals has done a great deal to promote Shareware for PC’s and
something similar on a smaller scale dedicated to promoting and
improving the quality of PD and Shareware for the Archimedes would be a
good thing.
6.3
I don’t mean a cosy club or an ideas forum but a genuine association
whose members would be obliged to conform to certain standards, whether
for Shareware or PD. If anyone thinks that this would be a good thing
then please write to me and I will try to put any interested parties in
touch with one another. What I won’t do, as has been suggested, is to
actually form an association because something like this should be
formed by a group of equals and not by a single individual who then asks
others to join. If enough people think it’s worthwhile then I would
certainly wish to be actively involved and will do everything that I can
to help.
6.3
Like the A.S.P. it would probably also involve participation by
libraries so I am also hoping for comment from them.
6.3
Please keep writing to me at the usual address − 39 Knighton Park Road,
Sydenham, London SE26 5RN. A
6.3
Dalriada
6.3
New artwork
6.3
Look Systems
6.3
From 6.2 page 5
6.3
Spacetech
6.3
From 6.2 page 16
6.3
Hardware Column
6.3
Brian Cowan
6.3
Sound from the PC cards
6.3
In a previous issue, I mentioned that that the Aleph One PC card was
capable of producing correct PC sound when connected internally to the
host Archimedes’ audio circuitry. This prompted Philip Draper to write
in and ask how to make the connections. Information is available in the
Sound Application Note from Aleph One, but for those interested, I
include some more details.
6.3
You need two 20cm lengths of wire, preferably different colours, with
0.1“ female sockets at each end, looking something like this:
6.3
6.3
One end will plug into the sound output pins of the PC card as shown
below. Be sure to note which is the “hot” line and which is ground.
6.3
6.3
The other end of the cable plugs into the computer’s “Auxiliary Audio
Connector”. This consists of two rows of five pins located on the main
circuit board towards the rear of the computer, close to the headphone
socket. The connector has different identification numbers for different
Archimedes models. On the 440 it is labelled PL2, as it is on the 410/1
range. However, the set of pins is denoted by LK9 on the 540 machines
and LK3 on the A5000 models. Also, note that the order of the numbering
of the pins is reversed on the A5000. Pin 5 is the hot line and pin 6 is
the ground.
6.3
I have indicated on the accompanying diagrams the location of the
connector for these four models; unfortunately I have been unable to
look inside any of the others.
6.3
6.3
Make sure that the ground line from the PC card goes to the ground line
(pin 6) of the computer board and the output line from the PC card goes
to the input line (pin 5) of the computer board. If everything is
connected correctly, your Archimedes will be capable of producing good
quality PC sound. This may be tested using the sound commands of a DOS
version of Basic or by playing one of the many PC games. I have found
the sounds produced by some games quite impressive − particularly when
compared with the tuneless grunts produced by the PC Emulator.
6.3
I have also indicated on the PC card diagram the arrangement of the
links for the RAM configuration. There are three links present on the
board when 1Mb of RAM is used; the diagram shows this. If you are
upgrading to 4Mb (by exchanging the eight RAM chips for 1Mb×4 80ns ZIP
DRAMs), the middle link must be removed.
6.3
387 chips
6.3
Philip mentioned that there was a (relatively) cheap 387SX chip
available from a firm called Datrontech in Aldershot, manufactured by
ULSI (sic) in California. He says that the chip comes with a disc to
demonstrate how fast it goes, and that the improvements are impressive,
although not quite as large as to a “real” PC. My guess is that the
video emulation is still the bottleneck.
6.3
Incidentally, I know that these third-party manufacturers of PC chips
are causing somewhat of a headache for the bods at Intel; they are
losing quite a lot of business. I have a relative working in Intel’s
future products department where what we would call the ‘586 (they call
it Pentium, a name which can be protected) and the ’686 are being
developed. He explained to me that the pirates certainly cannot copy the
silicon design of the chips: this is copyright. Thus while the cheap
chips should implement all published behaviour, there may well be
inconsistencies in some of the undocumented features. I know that some
people have had problems using 387 chips from particular manufacturers
(on real PCs) and this is the explanation.
6.3
Coprocessor algorithms
6.3
While we are on the topic of 387 chips, I would like to use this column
for a request for information. Some time ago, I read in Byte magazine
about the Cyrix range of maths coprocessors for the PC. As I understand
it, these are marketed as 387, or whatever, replacements, but the
internal mathematics is performed in a revolutionary way so that
calculations of transcendentals, for instance, are done very much
faster. Does anyone know in what way the Cyrix algorithms differ from
the more conventional ones? The description in the Byte article seemed
equally applicable to the traditional coprocessor chips. Perhaps it is
just a question of hard-wiring versus microcode − I don’t know. Any
ideas?
6.3
Quality graphics
6.3
Philip goes on to make one final comment in his letter, taking me to
task for doing the latest version of the PC emulator “a bit of an
injustice” in relation to the VGA colour implementation. As the new,
enlarged emulator manual says, VGA is fully emulated. Some of the 256
colour displays can be most impressive.
6.3
Emulator future features
6.3
In one way, though, the emulator is still lacking. As Mike Harrison has
pointed out to me, there is still no support for PC software that drives
the printer and serial ports directly, even on machines with the ‘new’
hardware, which is exactly what the PC software is expecting to drive
(so no emulation is required). This would be almost trivial to implement
and it would enable one to use such things as PC dongles. This must be
implemented in future versions of the emulator. Mike would then be able
to use his Seiko label printer on the A4, but this omission prevents it
working.
6.3
Hopefully, soon there will be a ’286 emulation so that Windows 3.1 can
be run.
6.3
’486 cards
6.3
Aleph One are now also producing PC486 cards. As with the ’386, these
come with a choice of one or 4Mb of RAM. The CPU is a Cyrix ’486SLC
which is pin-for-pin compatible with the ’386SX. However, as with a
regular ’486, it has a 1Kb on-chip RAM cache and a few more instruc
tions, specifically related to the cache. Unlike the full-blown ’486
chip, however, but like the ’486SX, there is no built-in floating point
coprocessor. Also, as with the ’386PC card, there is a socket on the
board for this. I am fortunate enough to have one of these cards but I
await the arrival of the FPU before installing it. I understand that the
performance should match a ’386 running at 50MHz! With the soon-to-be-
released Windows drivers, this should be quite an impressive product.
There is also CD-ROM support provided with the software.
6.3
RAM waste
6.3
I now have 4Mb of RAM in my PC386 card. I need to have this amount of
RAM (the maximum the card will support) so that I can run some large
Windows applications. So far, I have not quite got them running yet but
one thing is patently clear: it is a gross extravagance to have 4Mb of
RAM tied up solely in the service of the PC card. What is needed is for
some well-meaning individual to write the code to use this memory as a
RAM disk when the ’386 CPU is not being used. I believe something along
these lines was done for the BBC Master 512 card although I am not sure
if it ever saw the light of day. There is the complication of what
happens to the “disk” if the ’386 is subsequently turned on, but I am
sure that could be sorted out. If such a software application were
available, I am sure the 4Mb option would be more attractive for many
users.
6.3
RAM as virtual RAM
6.3
There is also a converse side to this argument. Windows makes extensive
use of virtual memory, by swapping chunks of data between RAM and disc.
So, fundamentally, you should never require more than 4Mb of PC RAM
(which is the limit on the present Aleph One PC cards). However, the
swapping will slow the system down. A solution is to configure one of
the “hard disc partitions” to be the Archimedes RAM filing system. Of
course, this is slower than direct on-card RAM but it is faster than
hard disc operations.
6.3
There are a number of limitations to this scheme however. Firstly, the
maximum size of RAM disc allowed by RISC-OS is 4Mb. Secondly, for some
reason the DOS FDISK command recognises that it is not looking at a real
disc and so it will not initialize the partition. What you must do is to
create a 4Mb partition on disc and initialise and format it before
copying it to RAM. Thirdly, there is the limitation of two hard disc
partitions so now you only have one left. However, if the RAM disc is
drive C, containing the DOS files, then the operation is blindingly
fast.
6.3
Ideally, a compressed version of the drive C (containing only DOS and
the empty swap area) could be stored in the PC application. This could
then be decompressed and loaded into the newly created RISC-OS RAM
filing system as part of the PC card’s boot sequence.
6.3
Also, such a scheme would provide some means of virus protection since
the DOS operating system is always freshly loaded from disc and the disc
need never be written to. In this way, the virtual RAM is provided by
real RAM from the host computer.
6.3
A4 report
6.3
So far, two readers have written to me reporting problems with their A4
machines hanging up in a manner vaguely similar to mine. Are there any
others out there? I suppose one consideration is that there are still
not many A4s around. Funnily enough, I have not had a hang for some time
(except on one occasion when the machine would not turn on). The only
explanation I can think of is that, in my initial euphoria, I was using
the A4 in the bath, (Archimedes would be proud of you, Brian! Ed.) in
traffic jams driving to work and in boring meetings. However, having
seen how rapidly the battery discharges, I now almost always use it
connected to the mains − using my long extension power lead. Maybe hang-
ups only occur with battery-only operation.
6.3
One reader found that his machine seemed to freeze after turning it on.
The other reader stated that his problems were particularly when using
the floppy disc drive with the PC emulator. His problems went away when
he switched off interactive file copying.
6.3
Portable patches
6.3
Quite a few of the common complaints concerning features of the A4 can
be alleviated with small software patches, often using relocatable
modules or sprite files. One may, in this way, change the functions of
various keys, alter pointers, etc. In the September issue of Archive,
Paul mentioned that he had created an enlarged screen pointer and
blackened up the Impression, pointer caret. I am sure other A4 users
must have written code for these machines. If there is demand, I will
collate any offerings to produce an A4 utilities shareware or careware
disc. Mike Harrison has made the point to me that it would be more
sensible if the display blanking, and the hard disc spin-down, were
inhibited when external power is plugged in. He might produce some code
to do this.
6.3
I have created an Impression function key strip for the A4 and I include
it on this month’s magazine disc. This is an Impression document and
readers will be able to edit it to create function strips for their
favourite applications. It is quite wide and utilises the facility of my
Laser Direct printer to print quite close to the edge of the paper −
beware!
6.3
CPU choice
6.3
I had been wondering why Acorn did not use the ARM 250 in their portable
machine. It seemed to me that that would provide the ultimate in size
reduction. Certainly performance would be reduced and expansion options
would be severely limited (although no one so far seems prepared to make
either an 8Mb RAM upgrade or a SCSI interface). However, there are also
VIDC considerations. Apparently the ARM 250‘s internal VIDC does not
provide access to the extra video lines necessary to drive the port
able’s LCD display.
6.3
New VIDC launched
6.3
Talking of VIDCs, the much-rumoured new VIDC chip has been officially
launched by ARM Ltd. It is known as the VIDC20 and Acorn have stated
that they will be using it in “future machines”. One wonders how this
will sit in relation to the various video enhancer boards now appearing.
I just hope that the owners of the older machines do not get left
behind. Computer Concepts have the right idea with Artworks (and
presumably Impression) supporting 24 bit colour, but not demanding 24
bit hardware to be present. (See press releases below. Ed.)
6.3
Sinister genius
6.3
I was raving, in the last issue, about the Genius tracker ball as an
alternative to the common mouse. I explained that Atomwide have rewired
the connector for connection to the A4 and that the assembly clips
nicely on the right hand side of the A4‘s case. Of course, left-handed
people (of whom there are a larger proportion than average in the
computer world) would clip the thing to the left side of the case. They
would then find that the lead was not long enough, the pointer moved in
the wrong direction and <select> and <adjust> were interchanged.
6.3
Never let it be said that there is prejudice in the computer world
against minorities; there is now a special version of the Genius for
left-handed people which has a longer lead specially wired so that the
pointer moves in the correct directions and the buttons have their
expected functions. This is quite a sensible thing to do and the
sinister (Latin for left − sorry, my joke!) Genius is easily distin
guished by its longer tail.
6.3
Incidentally, I did not seriously consider using the other tracker ball
(the one from PEP Associates) as this one uses the serial port. That
would use special software, it would tie up the serial port (probably
most required for a modem) and it just seems a complicated way of
solving a simple problem. After all, if the mouse port is provided, one
may as well use it.
6.3
FPU news
6.3
The latest news is that Acorn’s FPU chip should be released in the
“second quarter” of 1993. I understand the tape-up is completed and that
ARM Ltd will be commencing manufacture soon. First samples should be
available by the end of this year − but not to the public.
6.3
ARM3 + FPU upgrades
6.3
In a recent Hardware Column, I set out a challenge: who would be the
first company to sell an ARM3 upgrade board for the older machines which
included a socket for the FPU. Well, Simtec Electronics (0772−812863)
have come up with the goods, and at a reasonable price: £175 plus VAT.
They are selling an upgrade board containing a 25MHz ARM3 with an empty
socket so that when the FPU appears, it can be plugged in directly. It
is a beautifully made product with a proper plug (as on the Aleph One
ARM3 upgrade) which fits into the vacated ARM2 socket. The acid test
will have to wait for the appearance of the FPA chip although, of
course, then one will need the relevant software to support it. However,
I am not sure what the licensing situation for the FPA support software
will be. My guess is that the actual chip will cost between £75 and £150
but one may well have to purchase the software from Acorn (in the form
of a cut-down floating point emulator) on top of this.
6.3
Users of Acorn’s old floating point coprocessor card will recall that it
was necessary to provide software support for the floating point
operations which the actual FP chip still did not do. I understand that
the new FPU chip is remarkable in a number of ways. On the positive
side, the speed increase in its floating point operations is supposed to
be quite staggering − of the order of 1,000 times. On the negative side,
all it does is the elementary operations of addition, subtraction,
multiplication and division. Transcendentals and scaling of numerical
arguments to their proper interval are still done with the software
support by splitting them up into the elementary operations and passing
them to the hardware.
6.3
SCSI cards
6.3
There is an increasing selection of SCSI cards available for the
Archimedes owner. I have experience of using three: the standard Acorn
card, Oak’s SCSI card and the model produced by Lindis. The adverts
emphasise two main features: speed and the possible existence of an on-
board RAM cache. However, from experience, there is another considera
tion of importance, namely the software and the implementation of the
SCSI filing system.
6.3
My A540 came with an Acorn SCSI card. This functioned very well − after
all, the Acorn card sets the standard. However, I was disappointed with
with various features of the Acorn interface when using SyQuest
removable discs. Firstly, there was no real support for removable hard
discs and secondly, the various formatting programs laid down slightly
different boot sectors. In particular, the Oak card’s formatter writes
information of particular use to removable discs. I found the Oak
interface so much more convenient in use that I don’t use the Acorn card
any more. The lesson here is that the software/firmware is important to
consider when deciding which card to purchase. It should, however, be
emphasised that only the Acorn card (to the best of my knowledge) can be
used if one is running Unix.
6.3
Although I have not seen it, there is now an interface card from Arxe
Systems which combines a SCSI interface with support for a high density
floppy disc drive. I am sure that the hardware is superb; what remains
to be seen is how its software functions.
6.3
High density floppy interface
6.3
Last month, I discussed that there were rumours of a high density floppy
disc drive interface which was sufficiently small, and cleverly
designed, that it did not take up a podule slot. Michael Lowe has
written in to say that he saw the advertisement in the April 1992 issue
of Acorn User. It was from a firm called Evolution Computer Ltd,
announcing a HD interface and drive for £85. Unfortunately, Michael has
been unable to contact the company at either of its advertised
addresses. Do any of our readers have any further information? (The
address of Evolution and its UK arm, MAPS Ltd, are in the Factfile at
the back of the magazine. Ed.)
6.3
The bidirectional printer port
6.3
Regular readers of this column will know that I have been extolling the
virtues of the PC compatible bidirectional parallel printer port. I
mentioned in the last issue the range of peripheral devices in the PC
world which are available for connecting to such a port. I omitted to
mention tape streamers and I have now discovered hard disc drives and
even floppy drives which can be connected directly to this port. On a
PC, one of these devices would appear as an ordinary “drive” and the
software is designed to keep things as simple as this.
6.3
Mike Harrison has been looking into the software situation from the
Archimedes end, for using the printer port of RISC-OS machines with the
new hardware, as I mentioned above, in connection with the PC emulator.
He was rather disappointed by the apparent lack of software support in
RISC-OS 3.1 for driving the PC type interfaces at anything other than
the standard addresses. The only obvious provision for this seems to be
for an alternative IDE controller.
6.3
RISC-OS 3.1 upgrades
6.3
The operating system upgrades are now available and there are various
versions. The cheapest is the A5000 upgrade kit, and readers may be
wondering whether they can use the chips from that kit on the other
Archimedes machines. The answer is that you could in theory but by doing
so, you would be specifically breaking Acorn’s licensing agreement − and
it would be taking unfair advantage of them. The £20 you pay for the
A5000 upgrade is for the upgrade from RISC-OS 3.0 to RISC-OS 3.10 and
NOT from RISC-OS 2.0 to RISC-OS 3.10. Using the A5000 version on other
computers would be akin to borrowing a copy of version 2.10 of Impres
sion in order to get CC to give you a version 2.17 when you had only
ever bought Impression version 1.
6.3
ARM 3 upgrades and RISC-OS 3
6.3
Under RISC-OS 2, if you installed an ARM 3 upgrade to one of the ARM 2
machines you needed to run a special software application to turn the
cache on. This is all taken care of with RISC-OS 3 which includes cache
control in the operating system. I found my various Aleph One ARM 3
upgrades and the one from Simtec (mentioned above) all functioned
perfectly when their machines were upgraded to RISC-OS 3. A
6.3
A press release from Acorn about the new VIDC chip... Acorn has welcomed
the announcement from Advanced RISC Machines Ltd of its VIDC20 video
controller which has been designed to meet the needs of a wide range of
video systems, from power-sensitive portable computers through to
desktop multimedia workstations and from graphics accelerators to games
systems.
6.3
This versatile controller can display at up to 16 million colours at
VGA, Super VGA or XGA levels of resolution. Its low power architecture
will drive the LCD screen of a portable computer directly whilst
preserving battery life. It can be used to generate up to 32 bits per
pixel at data rates of up to 160Mb/s.
6.3
As well as providing a programmable display system, the chip also has
sophisticated audio support. The device is fully compatible with the
previous VIDC capabilities, providing up to eight channels of stereo
sound. For higher quality multi-media applications requiring both sound
and vision, the design provides a straightforward interface to dedicated
sound controllers.
6.3
Acorn Computers’ Technical Director, Malcolm Bird confirmed that Acorn
was fully involved in the specification for ARM’s new graphics control
ler with a view to it being used in future products. “The VIDC20 is one
in a new series of high performance chips from ARM Ltd that keeps the
ARM component set at the forefront of cost effective technology. We are
leading the test validation programme for the device and anticipate
including it in future generations of computers”, he said.
6.3
Here are parts of a press release from ARM Ltd that give more technical
details about the new VIDC chip... ...the VIDC20 is a high performance
video controller that consumes very little power (typically 0.3W)...
...other features include a hardware cursor, programmable pixel rates
and an on-chip sound system.
6.3
VIDC20‘s low power architecture will drive the LCD screen of a portable
computer directly, whilst preserving battery life, using a patented 16-
level grey scale algorithm to give high quality picture rendition.
However, VIDC20 is also capable of directly driving a colour CRT,
allowing the same portable computer to use a colour monitor on the desk
without additional components. The power management techniques ensure
that functions not in use, such as the video digital to analogue
converters (DAC) and sound DACs do not consume power. In addition, the
on-board palette has been segmented so that only one eighth is enabled
at any one time.
6.3
VIDC20 was designed with the needs for higher levels of display
resolution in mind. The chip can be used to generate 1, 2, 4, 8, 16 and
32-bits/pixel formats at data rates of up to 160Mb/s. The on-board 8-bit
linear DACs give a total of 16 million possible colours at rates of up
to 100MHz and drive doubly terminated 75R lines directly. The inclusion
of an on-chip phase comparator, when used with a voltage controlled
oscillator (VCO), allows a single reference clock to generate all of the
frequencies for any display mode.
6.3
VIDC20 also supports a hardware cursor in all of its modes − high
resolution, interlace and LCD. By offering cursor support on-chip, the
designer benefits from higher performance and lower software overhead
resulting in a better “look and feel” for the user. The cursor is 32
pixels wide, an unlimited number of pixels high and can be displayed in
4 colours from its own 28-bit wide palette.
6.3
As well as providing a programmable display system, the chip also
supports two different sound systems. An on-chip system incorporates an
8-bit µ-law DAC providing up to eight channels of stereo sound. A 32-bit
serial sound interface output enables the connection of external CD
DACs. The on-chip sound system makes the development of powerful, yet
low chip-count, X-terminal controllers straightforward. For higher
quality multi-media applications requiring both sound and vision, the
design provides a straight-forward interface. A
6.3
ICS
6.3
From 6.2 page 32
6.3
Language Column
6.3
David Wild
6.3
I was very interested to see Gerald Fitton’s remarks, in the last issue
of Archive, about the vast reduction in the amount of time spent
programming these days. Partly, I suppose, this is because software
packages have become much “cleverer” and many of the tasks for which we
would once have written programs can now be done inside a package. This
is probably one of the reasons why the language ‘APL’ didn’t take off in
the way that many of its supporters hoped. Many, though not all, of the
clever facilities for manipulating blocks of data are available in
practically every spreadsheet there is.
6.3
I have written several Turbo Pascal programs at work recently but all
they did was to take the report output from a database package and alter
the format to correspond with that required by the mainframe program to
which the file has to be sent. Although the programs have quite a lot of
detail in them, the only computation in the old sense is the production
of record counts and file totals.
6.3
Another problem these days, especially with the rise of WIMP usage, is
the amount of work required just to be able to accept some data and
print the results. With compiled languages such as ‘C’ and Pascal, there
may be very much more programming in providing a test environment for a
function than there is for the function itself and this can be a
disincentive if there is no compulsion to do the work.
6.3
Something that can help to restore your appetite for programming
experiments is a language like Scheme which is a modern version of LISP.
Ray Burcham, who is in charge of distribution of Scheme in this country,
has sent me a newsletter which says that a WIMP version of Scheme will
be available for the Archimedes in February, and it will cost only £5
more than the present single-tasking version. As this costs less than
£50, it is a very reasonable way of adding to your repertoire. While it
is not a language in which you can write programs for distribution, it
is a marvellous environment for developing algorithms. You can modify
functions “on the fly” and test them by just typing a call with the
appropriate data.
6.3
A recent article in one of the computing weeklies talked about prototyp
ing. The author warned that, although it had many advantages in showing
users what could be done to respond to their needs, it was very
important to ensure that sloppy code from this stage did not get through
to the production phase. A language like Scheme could be very helpful
for this as the algorithms would have to be recoded in one of the
compiled languages anyway.
6.3
One or two readers have already responded to my plea for help in my last
article. Being a ‘C’ novice, I had not appreciated the way the ‘scanf’
statement was specific to a data type, and I am very grateful for the
explanations.
6.3
Michael Kinghan of Oxford wrote a letter, which Paul has put in the
Comment Column on page 44, taking me to task over several aspects of my
article. I must apologise for the typing mistakes in the example program
and can only say that this was caused by typing it quickly without
sufficient checking. When I tried the program I didn’t think that it was
worth keeping and it was only when I came to write the article that I
realised that I could ask you for help and so had to retype it.
6.3
I still don’t understand why the ‘C’ libraries need to be at root level,
and this would still be the case if they were Pascal libraries. Except
for very occasional updating, there seems to be no need for me to get at
them and I would rather they didn’t clutter up the desktop when they are
not needed. What I would have liked was for the installation routine to
ask me where I wanted them to be put, and I would have said in the
directory “languages” with something like the path statements used by
programs like Impression to set the correct root.
6.3
There was no particular disappointment at not making the ‘C’ routines
work with Pascal ‘string’ types; they did work with variables defined as
‘packed array of char’ so I was quite happy with that − perhaps I did
something wrong and I’ll try again sometime.
6.3
It was not the ‘Plib’ library which I dragged in with the file to be
compiled but my own library, and this turned out to be quite convenient
as I could drag ‘aof’ files as well without creating libraries. I
certainly didn’t alter the value of ‘P:’ as the ‘Plib’ library was
linked in again as soon as I took my own library out of the libraries
list. This is another area for experiment but I did get my program
compiled and working satisfactorily.
6.3
I would still like to hear of a method for initialising static vari
ables, and details of any other problems such as some of the ‘ISO-
standard’ functions not working if extensions are enabled. I certainly
like the new compiler and hope that someone will write a proper manual
for it eventually. A
6.3
Comment Column
6.3
• Acorn’s marketing policy − I was interested to see the comments by G T
Swain and Ian Lynch (Archive 6.1 p16).
6.3
I, too, get a little tired of the advertised discounts for education. No
one minds a discount for quantity but your current advert for teachers
of a £200 discount on the A4 I find distasteful. Ian quite rightly
points out enthusiasts are an important market niche but they must not
be insulted. They have to use their own hard cash which comes from an
overtaxed income.
6.3
MSDOS. Acorn ignore this at their peril. When I bought my Archimedes, it
represented good value at the price, compared with the PC market. A
small premium on the price was understood and accepted. Today this has
all changed. With the price of a ’486 having fallen to a quarter of what
it was, the competition is fierce and the premium is becoming too high.
6.3
Once upon a time, everyone I worked with had BBC model B’s and then
Masters. The company I work for, a major airline, is working hard to get
everyone computer-minded. Needless to say, all the seminars that the
company have given revolve around PC machines and Windows. Of all the
people I work with, only three that I know of, have Archimedes; all the
rest have junked their BBC’s and bought ’386s or ’486s.
6.3
Many of my colleagues ask me what to buy. What am I to say? We have a
requirement to run some in-house PC programs and the industry standard
is rapidly becoming a ’486 with DOS 5 and Windows 3.1.
6.3
‘Windows is awful’, I hear everyone cry! You know that and I know that
but try telling someone who doesn’t know what a window is. He wants the
comfort of major software packages and something that is ‘standard’. Let
us not delude ourselves.
6.3
The PC Emulator. If you need to fiddle with some text processing on the
side this is fine. As I have said, I have to run an in-house PC program
every month, which takes 15 to 20 minutes to give a result. Every time I
alter the data, even slightly, it’s another 20 minute wait. A slow ’386
takes 3minutes, a ’486 barely takes 30 seconds.
6.3
‘Buy a PC’, I hear Ian say. This is fine and certainly cheaper than the
Aleph expansion card, (even after the price reduction) but what next?
Microsoft are not standing still. They know the shortcomings of their
software and it will change. If I bought a ’486, I would end up selling
my Archimedes and it will be one less customer for Acorn.
6.3
A4 Notebook. I would dearly love to have one, as I could make a great
deal of use of it. However, I can buy a good spec ’486 and a good spec
’386 notebook for almost the same price as the A4. You can see why I
think the premium is getting too high. How can a non-business enthusiast
justify another £2,000, not to mention all the insurance costs, for a
second computer?
6.3
Finally, the headmaster of my son’s school pointed out to me that, as
IBM PC’s are the industry standard, that is what is used in his school.
It is hard to dismiss this logic, particularly with today’s prices of
PC’s.
6.3
If Acorn junked the PC Emulator and supplied the A5000 with an Aleph One
’386 expansion board as standard, the Archimedes would become exceed
ingly attractive. Eddie Lord, West Sussex.
6.3
• Battery changing − I recently changed the batteries on my A420 − they
had been in for two years. When I switched on, the hard disc would not
work − the LED just came on and stayed on. In the end, I discovered that
firming in the HD data cables brought it back to life. I tested the old
batteries which gave 5.0A short circuit current and the battery tester
declared their condition as ‘good’!
6.3
The moral of the story? Well, firstly, I don’t think it is necessary to
change the batteries every year as suggested in Archive. Secondly, when
things go wrong, try the obvious things first, like firming the cables.
G Wilcockson, Saffron Walden.
6.3
• Error messages − Having seen the amusing error message in Archive 6.2
page 8, I thought I might as well send you this one: ‘Warning! Stray
debugging code detected. This machine will self-destruct in ten seconds.
Awooooga! Awooooga!’.
6.3
I found it in the Impression II file ‘Impress. Resources.uk’ and I have
only been able to make it appear by altering the error message code to
something like ‘InTi:’ and then selecting ‘Document − Info...’ from the
main menu. Note, you will have to restart Impression for it to load the
new message file ... Well I thought it was amusing... Philip Lardner,
Co. Wicklow, Ireland.
6.3
• Flopticals − As a back up medium, these seem ideal to me, particularly
as the prices will fall. (Discs are already down to £15 each). I chatted
to the man from Morley, who told me that floptical drives have a piece
of code written into the drive software that prevents them being used
without a software key − built into the SCSI board. Can I use an Acorn
SCSI board? − “no” Morley said. Why not? “Talk to Acorn” they said.
After much to’ing and fro’ing I got nowhere until I talked to Hugo of
Serial Port fame.
6.3
He confirmed the software lock problem and promptly showed me a
Floptical working on an A540 with Acorn SCSI board. Serial Port provide
a program that does the unlocking and allows the use of the Acorn SCSI
board. They indicated that this is available separately so that the
Floptical can be sourced elsewhere.
6.3
I look forward to next year’s Acorn User Show − definitely worth a
visit. Eddie Lord, W. Sussex.
6.3
• Impression Junior upgrades − Trick or Treat? Archive has stated (e.g.
4.2 p33) that Computer Concepts will upgrade Impression Junior to
Impression 2 “for the difference in cost”. Please note that this is not
the £90 difference in Archive’s advertised prices but £117.50, the
difference between Computer Concepts’ list prices. I probably would not
have purchased Impression Junior if I had known this to be the case, and
feel tricked into forking out an extra £27.50 needlessly. Instead, I
probably would have gone for the full package in the first place. I had
been contemplating this for a while before Impression Junior came along,
only being put off by the unpalatable dongle.
6.3
Having said that, Impression 2 is a treat to work with and I am pleased
that I have the full version. Impression Junior lacked just the features
I wanted. Am I overly suspicious in asking myself whether this was by
chance or design? John Savage, Leighton Buzzard.
6.3
(I would assume it was by design. If I were supplying a cut-down version
at a cut-down price, I would keep the best features for the full
version, wouldn’t you? Ed.)
6.3
• Language Column − David Wild in his last two Language Columns has
appeared to be having some aggravation with Desktop C and Desktop
Pascal.
6.3
October issue: David was disgruntled with the Desktop C installation for
placing a number of directories under $, rather than somewhere more out
of the way: “...for instance... the ‘Clib’ directory, to which I am
hardly ever going to need access...”
6.3
It may help him to reflect that his position is atypical. He has bought
a £240 C Development platform in order to write Pascal programs. One can
understand why he would like the C compiler’s resources to keep a low
profile, and also understand why they don’t! The Desktop C installation
creates four top level directories and their placement seems suitable
from a C programmer’s perspective. $.CLib, of which David specifically
complains, contains the Standard C Library (.o) and its headers (.h). To
C programmers, the Standard headers represent an online implementation
manual that is guaranteed to be true; so they like to have them handy (I
keep Clib on my Backdrop). The same goes for $.Risc_OSLib which contains
the RISC-OS Library and its headers.
6.3
David’s idea that the offending C directories might “hidden away
somewhere... with the correct paths maintained in... the DDE !Boot
application” is feasible and has been implemented. For the DDE C
compiler, the correct paths are maintained in (none other than) the DDE
!Boot application in the assignments to C$Path and C$Libroot. The
resources can be placed anywhere, as long as these variables point at
them. This works for DDE Pascal too. The only C resources it uses are
Stubs and Risc_OSLib and it looks for them on C$Path by default.
6.3
November issue: David could not get the Pascal compiler to find the Plib
library when he specified a library of his own in the Libraries option
of the compiler menu − not even when he specified Plib at the same time.
He was obliged to drag Plib onto the compiler dialogue. I have to report
that I am unable to reproduce this bug. Plib resides in <Pascal$Dir>.o,
and the link step finds it there automatically as long as P: remains set
to <Pascal$Dir> − and the !Pascal.!Run file sees to that. I have linked
my own development library, as well as Stubs, with Pascal programs via
the Libraries option without any hitches. I suggest that David has
accidentally or mistakenly assigned some path to P: in the course of his
experiments which does not, in fact, contain the Plib library. This
seems powerfully indicated by the fact that if P: is correctly set,
specifying P:o.Plib in the Libraries option, as David says he did,
causes that library to be included twice at linktime − because it is
included by default. The linker responds with a “duplicate file ignored”
warning which David would probably have mentioned had he seen it.
6.3
November issue: Disappointment with C Library functions that “are
defined to work on strings”. When called with Pascal string arguments,
the compiler threw a type mismatch error. The arguments had to be
redefined as packed arrays of char. I have not been able to reproduce
this difficulty either but, without David’s code, I cannot determine
just what he did. The following Pascal program calls two string
functions from the Standard C Library with Pascal string arguments.
Maybe it will make everything clear:
6.3
program bigears;
6.3
function strcmp(s1: string; s2: string) : integer; extern;
6.3
(* The Standard C Library function strcmp() compares s1 with s2
6.3
returning an int less than, equal to or greater than 0
6.3
according as s1 is less than, equal to or greater than s2 *)
6.3
6.3
procedure strcpy(s1: string; s2: string); extern;
6.3
(* The Standard C Library function strcpy() copies s1 to s2 *)
6.3
6.3
procedure compare(s1: string; s2: string);
6.3
begin
6.3
writeln(‘s1 = ’,s1);
6.3
writeln(‘s2 = ’,s2);
6.3
if strcmp(s1,s2) = 0 then writeln(‘s1 = s2’)
6.3
else writeln(‘s1 <> s2’);
6.3
end;
6.3
6.3
procedure noddy;
6.3
var s1, s2 : string;
6.3
begin
6.3
s1 := ‘Mike’;
6.3
s2 := ‘Kinghan’;
6.3
compare(s1,s2);
6.3
strcpy(s2,s1);
6.3
compare(s1,s2);
6.3
end;
6.3
6.3
begin
6.3
noddy;
6.3
end.
6.3
November issue: Problems with a little demonstration program from Donald
Alcock’s “Illustrating C”. The program is supposed to calculate interest
payments given a principal, an annual rate and a term. (a) Desktop C
died trying to compile it until calls to the power function pow() were
pulled from their original context. (b) Then it issued “lack of
precision” warnings but produced correct results. (c) Lastly, David
noticed that the power function is defined for double-float arguments,
while the program passes merely floats; but when he corrected this the
results were preposterously huge.
6.3
(a) This is a bug in the compiler’s elimination of common subexpres
sions. It needs to be reported to Acorn: I will leave that to David. To
explain, the compiler is trying to make good the inefficiency of Donald
Alcock’s code where he calculates the (very expensive) function
pow(1+R,N) twice within the same expression with identical values. The
compiler wants to hoist this calculation out and perform it just once
before evaluating the expression. Usually it succeeds but occasionally
it bites off more than it can chew.
6.3
There are two remedies. The efficient one is to do what David says he
did and eliminate the common subexpression by hand. The slack one is to
code #pragma no_optimise_cse at the top of the program. This just
disables common subexpression elimination. CSE optimisation does not
seem to be a strong point with the current C compiler. Back in the
spring, I reported another bug to Acorn that had the same cause. I had
been using the compiler for about 6 months before being bitten but it is
second nature to me not to code common subexpressions in the first place
because they are wantonly inefficient: I raised the bug by deliberating
using common subexpressions to detect signed integer overflow in certain
special cases. CSE problems very rarely cause the compiler to ‘abnormal
end’, as in David’s tale; but be grateful when they do! What I got was a
clean compile and a program whose behaviour was inscrutable nonsense.
Nonsense, moreover, that I could not reproduce in the debugger...
because the debug compile option silently disables CSE optimisation. To
anyone who doubts they have the strength for experiences like this, I
recommend they always enforce #pragma no_optimise _cse and rely on the
optimiser between their ears. Just one more bitter word on this topic:
the C manual (p378) prescribes the relevant directive as #pragma
nooptimise_cse. That is wrong: no_optimise_cse is right.
6.3
(b) The warnings do not signify lack of precision. They say “lower
precision in wider context <operator>”, and mean that an arithmetic or
logical context is forcing values to widen in order to balance the
operands to <operator> (i.e. to bring them to a common type). These
complaints are caused by Donald Alcock’s casual coding, in which both
integer constants and float variables are used as operands to double
float arithmetic. The warnings are saying: “If you really want wider
types at this point why don’t you say so?” If 12 and 1 are changed to
12.0 and 1.0 respectively they will be parsed as double float in the
first instance: declaring the float variables as double instead gets rid
of the remaining warnings.
6.3
(c) There is no need to ensure that arguments to the pow() function are
declared as double rather than float. The precision warnings make it
clear that this is not the problem. The C maths functions are all
defined for double float parameters and compilers automatically promote
float arguments to double. (The nuisance in fact lies in the difficulty
of doing single float maths when that is all you want.) The absurdly
huge results obtained by changing the float variables to double are
caused by neglecting the corresponding changes that need to be made to
the format specifiers in the scanf() function. These were left as “%f”,
which means “read a single precision (32-bit) float”, when what we now
want to read is a double precision (64-bit) float; so single precision
values are copied into the high-order words of double precision
variables giving garbage quantities that naturally tend to be big.
Change “%f” to “%lf” in the scanf() function − that means “read a double
float” − and the program is again correct.
6.3
A parting word to the editor: Can I stress the importance of correctly
transcribing program code for the magazine? Most of the folk who type in
listings are beginners and they can be baffled and frustrated by
transcription errors that they can’t understand. The dozen or so
statements in David’s program as published contain several errors that I
assume were introduced by transcription, since any of them would have
prevented David from compiling the program. There is no substitute for
getting program code in machine readable form and leaving it unedited
throughout the publication process. If you have to transcribe or edit
it, it really should be rekeyed and compiled from early proofs, and the
very least that should let you sleep at night is to have it proof-read
by a programmer in the relevant language. Here is a rendering that will
compile spotlessly and execute correctly IF YOU DON’T CHANGE IT:
6.3
#include <stdio.h>
6.3
#include <math.h>
6.3
6.3
int main(void)
6.3
{p , Rpct, R, M, power;
6.3
int N;
6.3
6.3
printf(“\nEnter: Principal, Rate, No. of yrs.\n”);
6.3
scanf(“%lf %lf %i”, &P, &Rpct, &N);
6.3
R=Rpct/100;
6.3
power=pow(1.00+R,N);
6.3
M=P*R*power/(12.00*(power-1.00));
6.3
printf(“\n£%1.2f,@%1.2f %% costs £%1.2f over %i years”,P,Rpct,M,N);
6.3
printf(“\nPayments will total £%1.2f\n”,12.00*M*N);
6.3
return 0;
6.3
}
6.3
Mike Kinghan, Oxford
6.3
6.3
(OK, I’ll try to do better next time. David Wild, in the Language Column
on page 42, accepts some of the blame, too. Ed.)
6.3
• Magic Moments by Richard Fallas... Paul has a problem. The weight of
Archive is (almost) literally pressing on his shoulders because of the
vast numbers of back-issues stored over his and his gallant staff’s
heads.
6.3
He is, understandably, concerned that the floor might not be capable of
providing continued support to all the unread pearls of wisdom accumu
lated above. Can we help? Yes! Buy back-issues. Can the Archimedes
(sorry, Acorn-acronymous-processing-machine) help? Possibly.
6.3
Let’s see if we can provide some reassurance by using the computer.
First of all, a bit of general background to what makes a floor stay up.
The floor is likely to be constructed of softwood joists (i.e. beams)
placed parallel to each other and spanning between two walls (let’s say
175mm × 50mm at 400mm centres spanning 3m). The capacity of the joists
is limited by several factors, e.g. shear and bending capacity,
deflection, etc. The most important two in domestic cases are likely to
be bending and deflection.
6.3
The job required of the joists is to carry the load above, but what
load? Well, there is the weight of the floor itself called the Dead
Load, likely to be of the order of 0.5 to 0.6 KN/m2. That reads “kilo
Newtons per square metre” − remember Isaac being bopped on the head by
an apple? A Newton is 100grammes x 9.81 m/s2 i.e. about ¼ lb, i.e. about
1 apple!
6.3
Next comes the Live Load, i.e. anything which can be removed. This has
to cover Archive staff, their furniture and the odd magazine or two they
leave lying around! In normal domestic cases, 1.5KN/m2 is used for
design.
6.3
A slight digression now: different types of building material are
considered in one of two ways for design, namely Safe and Ultimate. In
steel, for example, stresses used are those at the ultimate state, i.e.
when the beam buckles or bends in failure. Loads are factored up to
achieve an overall Safety Factor. Timber, by contrast, is currently
designed using Safe Working Stresses but with the actual loads expected.
The two methods give similar, but not identical, answers.
6.3
As this is a timber floor (we think) we will use the loads as they are
and restrict our stresses to “Safe” values. Here we have more variables
which affect our calculation: number of members acting together,
duration of load, type of wood, moisture content, etc, etc. Generally,
however, we can say that good softwood will safely sustain a bending
stress of 5.3N/mm2. This can be increased by 1.1 as we have several
joists acting together.
6.3
So to a design (or more accurately analysis)... Maximum Bending Moment
of a uniformly loaded beam is given by w × l2 / 8 (load per m times span
squared over 8). Bending Stress of a rectangular beam is given by
Bending Moment / Section modulus, z, where z = b × d2 / 6 (breadth times
depth squared over 6).
6.3
If you are still awake, you will see that we need to compare these
various elements in some way to see if Paul is now sitting with a pile
of rubble around him. Perhaps the best way is to calculate Moment of
Resistance and hope it is greater than Applied Moment. This is OK for
one case with all the variables known. If, however, we want to find what
the capacity is, with possible sizes, spacing and spans, etc, a
spreadsheet is an obvious way forward. Included on the monthly program
disc is a PipeDream template file which offers a means of altering the
variables and giving immediate results.
6.3
To pursue the one example, however, the sum goes as shown in the box at
the foot of the page.
6.3
If the values are as I have guessed, Paul now has a means of assessing
the risk of working downstairs. The template gives deflection values
which I won’t go into here. Of course, the example assumes uniform
loading and I know Paul has moved his magazines close to the walls. This
is too complex to deal with here but if he sends me a sketch with layout
and dimensions, I would be happy to check it further for him. The
template is able to cope with a distributed load which only applies to
part of the joist. Meanwhile, I hope this brief discourse into
Structural Engineering on the Archimedes has been of passing interest.
Anyone care to explore Nuclear Fusion or Relativity for us?
6.3
(That was an interesting aside − thanks, Richard. The template, plus
instructions, are on the monthly program disc. The magazines? Well, I
chickened out. I hired an industrial storage unit and moved the excess
magazines out. The removal company who did all the hard work reckoned
that they moved about 4½ tons down the stairs and out of the loading
bay. Now that’s what I call a weight off my mind!! Ed.)
6.3
• Pocket Book article addition − It would appear that there has been
some confusion about the hardware information in the Pocket Book article
from Archive 6.2 p61. The following is an excerpt from a letter by Chris
Parker:
6.3
“... the paragraph which discusses hardware differences seems to imply
that you can only use the 3-Link with an Archimedes if you run the PC
Emulator and make a few wiring changes.
6.3
“... you only need to make wiring changes and run the PC Emulator if you
want to run the...PC or Mac software supplied with the 3-Link. My Series
3 is connected to the serial port of my Archimedes with no wiring
changes. To transfer files from the Psion to the Archimedes, I use a PD
program called ‘DownLoad’... by Emmet Spier. To transfer files and OPL
Source Code from the Archimedes to the Psion I use the one-liner
Filer_OpenDir serial: cunningly disguised as an application.
6.3
“... the Psion does not appear as a filing system on the Archimedes but
files transfer in both directions...” Chris Parker
6.3
• Reviewing standards − I must say that I’m not too happy at the way in
which some of your review writers start their reviews. They often do not
clearly state what they are writing about. A typical example is the
review of Starch in Archive 6.2 page 64.
6.3
I may be teaching my grandma to suck eggs but the first paragraph of any
review should inform the reader of the name of the product, what it IS,
i.e. hardware or software, briefly what it is supposed to do, how much
it costs and who manufactures it. The second paragraph should outline
some of the ‘amazing’ features, or lack of, that make it such a
worthwhile buy, or why you wouldn’t touch it with a barge pole. It
should then go on in an informative fashion about the pros and cons of
the product and wind up with a brief conclusion which should summarise
the most important points made in the body of the review. It is the
editor’s job to cut out all the waffle and leave the reader in a
position to compare the product against similar ones. All too often,
Archive reviews fail miserably to do this and I for one do not have the
time to read badly written reviews. Perhaps you could give potential
reviewers an outline of how a review should be written, or print an
article in Archive to this purpose. Philip Lardner, Co. Wicklow,
Ireland.
6.3
Thanks, Philip. Your comments should help to point reviewers in the
right direction. The article on page 57 should also help. Ed. A
6.3
Actual Live Load (say) : 1 Archive = 120g approx. ( = 1.2N)
6.3
Area = 0.21 × 0.15
6.3
hence for 1m2, “Archive” Load = 1.2/(0.21 × 0.15) = 38 N/m2/mag.
6.3
if in piles 20 magazines high, Archive Load = 0.76 KN/m2
6.3
if in piles 40 high, Archive Load = 1.52 KN/m2 (i.e. our Limit)
6.3
Applied Moment = (1.52 + 0.6) × 0.4 (spacing) × 3 × 3 / 8 = 0.954 KN.m
6.3
Moment of Resistance = 5.3 × 1.1 × 50 × 175 × 175 / 6 × 106 = 1.48 KN.m
6.3
hence,..... M. o. R. > App. M. equals happiness.
6.3
Ovation Column
6.3
Maurice Edmundson
6.3
To my surprise, in the November issue of Archive a number of contribu
tors had cogent things to say about Ovation. Richard Hallas’ comments
were especially useful and many of the points he made are those I plan
to deal with more fully in this and later articles.
6.3
Text frames, picture frames and lines in Ovation are referred to as
“objects” and this option in the main menu leads to an object sub-menu
with a dozen tools for manipulating them in one way or another. The
labelling of these tools is automatically changed according to the
nature of the object selected. They give the user great flexibility when
planning a document.
6.3
However, before giving some examples of use, I want to refer to the View
option in the main menu and to the two choices Show Rulers and Units.
Usually you will probably find millimetres (mm) a suitable choice for
units, but you can also choose inches (“ or in) and points (pt). Ovation
uses units in an intelligent way. If you have globally selected mm with
the Units option but, at some stage, in your work, you require one
dimension in points, then by entering the numerical value followed by
pt, Ovation will correctly interpret this temporary special value in
points. Similarly, if you add ” or in after the number, it would assume
these to be inches. This is a great help in some situations and I will
give examples later. For now, I assume the Units are mm and Show Rulers
is selected so that on screen, there is a ruler marked in centimetres
and millimetres around the left and top borders of the document window.
6.3
If you are inserting a frame or line into the document, the ruler
assists in placing it at the correct position and also in drawing it to
the correct size. Once the frame tool is highlighted, short red markers
on the ruler show exactly where the mouse pointer is and hence where the
top left corner of the frame will be. As the mouse moves and the frame
is opened, these red indicators move and you can judge the width and
height of the frame quite accurately. Final adjustments can be made in
one of several ways. For accuracy to a fraction of a millimetre, not
only for the size of the frame but just where it is on the page, select
Modify Frame from the Object menu and write all the dimensions in the
dialogue box. If visual accuracy is sufficient, once the frames are on
the page and of a suitable size, with the mouse pointer over the frame,
clicking <adjust> will permit the frames to be slid very quickly about
the page into what you judge is the best layout. If <shift> is pressed
at the same time, the movement is constrained to the vertical or
horizontal. For these moves, the mouse pointer is replaced with a “move”
symbol.
6.3
The rulers also permit adding guidelines to the page and, providing the
feature is ticked, in the Object menu, frames and lines will snap to
guidelines when drawn close to them. (Used in the chess board example
below.)
6.3
I will make brief mention of a couple of special points about picture
frames which we will probably return to again in a later article. When a
sprite is loaded into a picture frame, with the pointer over the
picture, pressing <select> moves the picture within the frame but does
not move the frame. Draw a small text frame inside the picture frame (a
nested frame), prepare a suitable font and type in the caption to the
picture. With the pointer over the caption press <adjust> to position it
accurately; with the pointer over the picture, press <adjust> to move
the picture into position on the page. The nested caption moves with it
and retains its correct relative position. This method is better than
writing the caption externally beneath the picture.
6.3
The Modify Frame option also has a write-box for varying the angle of
the picture within the frame. Entering 70 would tip the picture
backwards by 70° and entering 320 would tip it forwards by 40°. Rotation
of sprites and draw files is only possible if you have RISC-OS 3.1 and
then may only be printed if your printer (such as directly driven
lasers) can cope with rotations.
6.3
If <shift> is pressed just before drawing a frame, it will automatically
be drawn as a square whatever size you wish to make it. If you start to
draw a frame and then feel you have started at the wrong point,
returning the mouse pointer to the point you started from and releasing
<select> will annul the frame (or press <Escape> at any position); the
frame can be started from another point immediately. A border (one of
many styles supplied or make up your own in !Paint) can be added to the
frame.
6.3
A very important tool in this versatile list, is the Duplicate Object
function. A worked example will best illustrate all the things I have
described. Consider the design of a chess board. Guidelines will also be
helpful in this exercise. They will assist with the final border around
the whole board. Open a document window with the rulers visible, and
move the mouse pointer on to the top ruler. Press <select> (or <adjust>)
and the vertical guideline appears. Slide this to the 20mm point and
release. Repeat along the side ruler to give a horizontal guideline 20mm
down. Snap to Guides in the Object menu should be ticked.
6.3
Draw a picture frame exactly 20 × 20mm where the guidelines intersect.
(Remember pressing <shift> forces a square). Select Duplicate Frame from
the Object menu. The following default dialogue box appears.
6.3
6.3
We are going to position all the white squares on the page. First create
the top line of white squares. The “number of times” will be 3 (we have
1 and require 3 more). We do not want any vertical offset but we want
the top left corner of each duplicate square to be positioned two square
widths away from its parent, i.e. its own width plus the width of the
intermediate black square. Therefore set horizontal offset to 40mm.
Click on OK and they appear as if by magic! Repeat for the left hand
vertical column. Here the vertical offset is 40mm and the horizontal
offset is 0mm. Now duplicate each of the top line squares downwards in
threes. Next duplicate the top left corner square once with 20mm
horizontal and 20mm vertical offset. Here is a view of a portion of the
screen at this stage.
6.3
6.3
Finally, duplicate this square in threes horizontally and vertically and
all the white squares are completed.
6.3
6.3
Returning to the top left corner, duplicate once with a 20mm horizontal
shift and then, from the Modify Frame dialogue box, change its colour to
black or grey.
6.3
This square is then duplicated in exactly the same way as the white
squares, to fill the chess board with the black squares. Because we have
used picture frames, if you have chess-piece sprites, they can be loaded
into the sq uares as required. Their size (scale) can be adjusted from
the Modify Frame dialogue box by changing the values in the two write
slots called X-scale and Y-scale. Now add guidelines exactly along the
right hand and lower edges of the board. so that the whole board is
framed with the guidelines. Draw one large picture frame around the
whole board, allowing it to snap to the guidelines. From Modify Frame
give it a border 2mm thick and make the frame transparent.
6.3
In cold print, all this may sound a little tedious and complicated but I
have put it forward as an exercise to illustrate the great versatility
of the Ovation frame system and, in practice, the whole chess board can
be drawn in a few minutes. From the View menu select Hide Guidelines to
check all is well before printing. This is a reduced view of part of the
printout from my ArcLaser.
6.3
6.3
The Duplicate Object function is ideal for producing all kinds of
tickets, membership cards, compliment slips and so forth. It is even
possible to produce numbered tickets, without too much trouble. Since we
are duplicating an object, not just a frame, everything contained within
the frame is duplicated as well. To assist with the final cutting of the
printed A4 sheets, it is best to choose ticket sizes which fit the sheet
exactly, not forgetting that outside margins of around 8 mm represent
the print limits for most printers.
6.3
I give an example of a numbered membership ticket to illustrate all the
techniques. I shall make it 105mm × 74.25mm which will give eight per A4
page and they will be very easy to guillotine along the major axes into
individual tickets. Use New Document to define a page with 8mm margins
all round. If you do not require numbered tickets, work on the Document
page. If you do require numbers, everything that follows is carried out
on the Master page. Near the top left corner draw a text frame about
85mm × 55mm a few mm inside the main frame. Add a 0.5mm border then
modify the frame to exactly X=8mm; Y=8mm; W=89mm; Ht=58.25mm. This will
not only make the frame the correct size but move it right into the
corner of the main frame. When borders are drawn around the outside of
frames, their dimensions are increased by the width of the border all
round. This must be taken into account when duplicating frames with
borders. In this instance, had the frame been placed in the corner
before the border was added, you would have been given an error message
saying that there was not enough room for the border. (It must remain
within the main frame.) Every ticket will have an 8mm wide margin all
round.
6.3
Now using suitable fonts, fill in the ticket with the required text.
Here is an example:
6.3
6.3
If numbering is not required, the frame is duplicated once to the right
(horizontal offset=105mm) and three times down (vertical
offset=74.25mm). Save it! then print as many pages as required to
complete the job. If numbering is required, proceed as follows. You have
to prepare a “numbers” document which, once done, can be used over and
over again. Open a new Ovation document with (say) 6 columns. Choose a
font and a style for the numbers e.g. 14pt. Homerton Bold. Starting at
the top of the first column type in numbers of three digits from 001 to
999 (or if two digits will suffice, from 01 to 99) with a carriage
return between each one. Hence we create a numbers list similar to the
following. (Save it!)
6.3
(How about using Basic to generate the numbers:
6.3
*Spool RAMFS:numbers
6.3
FOR N=1TO1000
6.3
PRINT N
6.3
NEXT
6.3
*SPOOL
6.3
Then load the file and do a bit of search & replace to get rid of
leading spaces and to add leading zeros for numbers less then 100. Ed.)
6.3
6.3
On the ticket, draw a small text frame where the number is to appear, of
such a size that it will just hold one of your three digit numbers. In
my sample, the number frame is 11mm × 8.25mm The corner ticket, which is
our template, is now complete. Duplicate it across and down as described
previously. Use the link tool to direct the text flow from number frame
to number frame starting top left, down the left hand tickets, up to the
top right and down to the bottom right. Return to the document page.
Delete the blank page to obtain the new page with all the tickets on it.
Save it! Put the caret anywhere in the main frame, then holding down
<return>, allow the document to accumulate as many pages as you need,
e.g. If 64 tickets, assemble 8 pages. Finally, starting with page 1, use
the link tool to link the bottom right number frame to the top left
number frame on page 2 and repeat this for all eight pages. From the
numbers document, use the Edit function to copy the first 64 numbers.
Then paste them in to the tiny frame on the top left ticket on page 1.
All the other 63 numbers will flow into the correct tickets. Print the
document. The above may sound rather long winded but it is surprising
how quickly a set of tickets for a modest size club can be printed. A
6.3
Fonts Workshop − Part 4
6.3
Roger Spooner
6.3
In the second article, Using FontEd, I described how to use Acorn’s font
editing program. I covered loading a font, moving points around,
creating and deleting things, live links, scaffold lines and so on.
However, there was little on how to design a pleasant looking character
set. I hope to cover that here.
6.3
Balance
6.3
Making a font where you can recognise the letters is easy. The problem
is how to make one which is clean and balanced − one which hangs
together well. This can be a very difficult task, not because of the
tools but because of the discerning eye of the reader.
6.3
To assist with this, though, it would be helpful to have some consistent
standards throughout the new font. A number of methods can be used,
including a framework of straight lines, a measurement of line thickness
and a comparison of widths.
6.3
A basic framework
6.3
To start with, it is important to design all the letters with a
consistent idea of scale and position. This is not easy on a plain white
background, so I do the following:
6.3
Think a little about your new font. Consider the typical height-width
ratio and the normal thickness of both horizontal and vertical lines.
6.3
Starting with a completely blank font window, open the skeleton window
for space. You will see a small red cross near the bottom left.
6.3
Create a horizontal scaffold line level with it and going up for the
thickness of a horizontal line segment. Do the same a bit to the right
of the red origin point (50 units) and set its thickness to the typical
vertical line thickness.
6.3
Create similar scaffold lines at about 300 along and 400 up. The precise
details are up to you.
6.3
These lines form the outside edge of the basic letter. It may be that
nothing is close enough to be worth linking properly but they help align
your work consistently.
6.3
Duplicating the framework
6.3
Once the scaffold lines are made, they should be copied (by dragging the
character box in the font window with <adjust> from one box to another)
so that the scaffold lines are inherited. Do not copy it everywhere; you
will want to add more scaffold lines to characters like B which can then
be copied from there to E and F. Be careful not to move any scaffold in
any character except the one where it was created. When you do move
scaffolds, all characters with copies of the same are also altered. If
you do have to move it, you may want to Make Local the line in question.
6.3
Designing the shapes
6.3
You can start designing actual letters immediately after setting up the
initial framework of scaffolds. You may choose to start on A, or perhaps
a simple letter like I, and then move to H, E, F, B, D, etc. It’s up to
you.
6.3
Be careful to keep the line thickness methodical, using the thickness of
the scaffold lines and perhaps using the shadow of another character
behind the one you are designing (drag its little box into the skeleton
window with <select>).
6.3
It is customary to make curves go out beyond the guiding lines. Thus an
O finishes below an I, and starts above it.
6.3
6.3
This is a very small variation, but it makes a world of difference.
Without it, curved characters look distinctly small by comparison to the
straight ones.
6.3
Regulating line thickness
6.3
It is difficult to get the line thickness right all the time. It is
certainly easiest when the line is horizontal or vertical, and so I
recommend that the end points of curves are at those limits. (Although
they are at 45° in the Acorn fonts, these were probably done by a
machine). Even with them at the top and bottom, you may still need to do
some arithmetic. This would involve turning on Display.Coords and
dragging a point nowhere. As you drag it, the coordinates (in design
units) are visible. With them, you can ensure that everything is how you
want it.
6.3
It is customary to make the lines thinner when horizontal. Where two
lines meet, it is sometimes necessary to dig into one to make the shape
distinct. The shape can be quite severely deformed before the reader is
actually bothered by it. This is shown in the ‘r’ of Homerton.Black,
where the curving part gets unusually thin as it joins the vertical.
6.3
6.3
This merely serves to accentuate the curved shape.
6.3
The good curve
6.3
Although you may be getting bored of it by now, it is worth pointing out
again what makes a good curve.
6.3
Imagine the crossing point of the lines from the end points to the
control points of the curve, as shown by the dotted line below.
6.3
6.3
If this crossing point is between one of the end points and its control
point (as in the diagram), the curve won’t look nice. The crossing point
should be beyond both control points.
6.3
As with contour lines on a map, outlines, be they curved or straight,
should never cross. Although it will be plotted correctly, you can be
assured that the shape is a compromise − you should really put more
effort into designing the shape carefully.
6.3
You should also keep the number of lines in a character to the minimum.
If you allow a program like Trace to do an outline of a scanned
character, it will probably come out with dozens of line sections. This
is not good for real fonts as it takes longer to plot, uses more disc
space and is more messy. Anyway, scanning should never be part of the
process of font design!
6.3
Live linking, the method for putting parts of several characters
together to form another like é, should be used wherever possible: a
colon should be a proper copy of a dot combined with a live link of a
dot (the latter moved up). An i should have a live linked dot (which
won’t get linked when you come to ï) and so on. You may want to do the
apostrophe without links, as that will save you from effort with double
inverted commas, and the polarised inverted commas.
6.3
Difficult characters
6.3
(I can think of a few! Ed.) Some characters are easy enough to create by
just drawing the outline immediately. The letter I, for example, is
often just a rectangle. With others like ‘&’ though, the complexity of
varying thickness and crossing lines means that you have a brain melt-
down while trying to figure out the outline positions and joins.
6.3
To solve this problem, you can do the skeleton lines first. These should
follow the exact shape of the letter (as compared to doing them later
when they just need to be inside the outline), including correct curves
and positions. Note that the skeleton runs through the middle of the
shape, so don’t go right to the edge of the intended outline.
6.3
You can then adjust the skeleton with ease before trying to put the
outline round it and can therefore be fairly sure that it’s right before
you go to too much trouble. Look at the red line in the Full Char window
to be sure of the shape.
6.3
Finally, just drop the outlines in, being careful to make them start and
end in places which will allow you to join them to lines going the other
way easily.
6.3
Widths
6.3
It is also important to keep the width of each character precise. Make a
stab at it, then save the file, load it in to something like KeyCaps
(from Design Concept) or Draw and take a long hard look at whether each
character looks well spaced compared with every other characters. Don’t
be afraid to go up to really big sizes like 1000 pt if needed − you’re
not on a PC now! If you do need to make an adjustment, remember it can
be to the left as well as to the right.
6.3
To change the space to the left of a character, select any scaffold
line, link all the character’s points to it (drag a big box round them
all), then hold down <Ctrl> and press <Left> or <Right> (¬ or ®). This
will move all the linked points 16 units in the given direction. Then,
press <F9> to unlink the points from the scaffold and relink the correct
points.
6.3
Because the width of the previous character specifies where the origin
of this one goes, you can control where the black actually starts by
moving the shape in relation to the origin.
6.3
Hopefully it will soon be possible to specify the gap between every pair
of characters (i.e. thousands of numbers) individually, using kerning
pairs. Thus you will be able to get the spacing exactly as you want it.
In the mean time, it’s not too difficult to create a spacing which suits
pretty well everyone. A
6.3
A Little OOPS
6.3
Jim Bailey
6.3
Smalltalk was originally both an operating system and a language; on
modern computers it can be considered just a language. It is an Object
Orientated Programming System (OOPS) and was developed at Xerox Palo
Alto Research Center during the 1970s. In 1980, a graphical front end
was developed resulting in Smalltalk 80 upon which current versions are
based.
6.3
The hardware required then was regarded as advanced; a 16 bit processor,
80 to 100Kb of memory, 5 or 10Mb hard disc, a high resolution mono
display and a three button mouse. At that time, an average home micro
had about 48Kb of memory and an 8 bit processor (6502, Z80) e.g Apple
II, PET and TRS80 Model I. The software consisted of 6 to 12Kb of
machine code to implement the interface to the hardware and to set up a
basic Smalltalk system. The bulk of Smalltalk, about 300Kb, was written
in Smalltalk itself. Modern Smalltalk systems can be very much larger.
6.3
Smalltalk is totally object orientated; everything in Smalltalk is an
object. Programming is done by defining Classes and creating Objects
from them. Classes become part of Smalltalk; thus programming actually
alters Smalltalk itself; Smalltalk is your program.
6.3
Classes define what an object can do. Most other languages are command
driven. In Basic, the command “PRINT item” calls a routine which
examines “item” and decides if it knows the type and how to print it; if
it’s not one of a limited number of types then PRINT fails. In Small
talk, the object “item” is sent the message “Print” and it is up to
“item” how it interprets the message. If the programmer is not happy
with the print method, he can rewrite “items” Class definition to his
own satisfaction; try that in Basic. All the Classes in Smalltalk can be
rewritten; but care must be taken when altering the basic Classes as an
error in these could make the system unuseable i.e. a total crash.
6.3
Variables in Smalltalk have only one type; they are all Objects. A
variable can start life as an integer, change to a string then become an
array. To become side-tracked for a moment − what is an array? An array
is usually a group of similar items such as strings but in Smalltalk,
only Objects exist, therefore only arrays of Objects can be created. An
Object can be anything that a Class defines; therefore an array of four
Objects can consist of an integer, a string, a sprite and another array.
6.3
Creating Classes can be done from scratch but the usual method is to
define a Class as a sub-class of an existing class which already does
most of what is needed and to define only the new methods needed. A sub-
class can do anything its parent Class(es) can do. A Class can be a sub-
class of a sub-class and it inherits all the methods of all its parent
classes.
6.3
Smalltalk for the Archimedes is expensive − over £400. The cheapest is
for DOS which costs under £100. For those who wish to try out OOPS,
there is a public domain version called Little Smalltalk. This was
written for UNIX-based text-only systems; and is written in C. It was
ported to the Archimedes several years ago. The version I have was
ported by Smalltalk Express to run under Arthur! I have converted it to
Standard C and it now runs as a normal application. It is still text-
based but I have added additional features to access more of RISC-OS.
6.3
Reference: A Little Smalltalk by Timothy Budd ISBN 0-201-10698-1. A
6.3
Mad Professor Mariarti
6.3
Stephen & David Sloan (aged 11 & 10)
6.3
This is a rather old game from Krisalis ported directly from the Amiga.
It is a platform and ladders game and comes in a small box, not unlike
that of Lemmings, with a cartoon-style picture on the front. It has a
couple of screenshots on the back, with the scenario in a few different
languages. Inside the box are the game disk and a small leaflet which
tells you the scenario again, the controls and the loading instructions
for the Amiga and Atari ST (groan!). The scenario goes along these
lines... You have built some laboratories without planning permission. A
few of your machines have turned mutant and have taken over the labs and
are running around outside frightening people. The mayor isn’t too happy
about this and wants you to shut down the labs − if you don’t he will
send Dr. Headbender to lock you up in the local asylum.
6.3
The game
6.3
At first when we inserted the disk with auto-boot configured, we got an
error message telling me that the boot file was for Arthur users only.
We returned to the desktop and loaded the game from there. When you load
up there is a screen almost matching that on the front of the box and
then the same picture with the word ™‰‘mad’ bouncing up and down. There is
also a good tune playing and if you don’t press space for a while, you
will come to a screen with instructions and credits and a better part of
the tune will play.
6.3
The keys are <Z> for left, <X> for right, <P> for up, <L> for down and
<space> for fire although they are fully redefinable by pressing <ctrl-
D> during the game. Pressing fire will bring you to the first game
screen where you can choose which laboratory you want to visit first.
There are five labs to be shut down: Chemical Research Lab, Computer
Science Lab, Space Rocket Development Lab, Biological Observations Lab
and the Mystery Lab. They can be shut down in any order except that the
Mystery Lab cannot be entered until the rest have been shut down. To
choose a lab you must get to its entrance by climbing, walking or
jumping and jump in the middle of it. There are no adversaries on this
screen but trying to walk off the edge of the screen will cause you to
be booted back on again.
6.3
The labs all have good cartoon-like graphics and excellent tunes but the
foreign language versions are not available on the Archimedes. To close
down the labs, you have to reach an almost inaccessible switch using
equipment to help you on your way through to it. You will need to get
certain items of equipment in order to get other bits of equipment that
enable you to shut down the lab. Some of the puzzles are quite humorous
− for example, there is one where you need to get some milk and some
cereal and put it in the microwave and then eat it to stop yourself from
freezing when you go outside onto the roof.
6.3
You have to fight your way through mutants using your trusty spanner to
destroy or freeze them so that you can get past − although they come
back after a while. There are different mutants on every level and they
all take varying amounts of shots depending on the weapon you have. They
look like ordinary objects that have been brought to life. For example,
you can find a mutant disk in the Computer Lab or a mutant calculator in
the Space Rocket Development Lab. You also have to watch where you jump
because there are certain things like electric sockets and boiling
liquids that kill you in a variety of ways.
6.3
Dotted around the labs are tokens that look like coins spinning round.
Once you have collected enough of these you can trade them in at a tool
vendor to get upgraded weapons. The weapons are listed here from least
to most effective: Spanners, Screwdrivers, Bolt-gun, Fireball-thrower,
Laser-gun and Virus-killer.
6.3
Problems
6.3
There are a few slightly irritating problems and one very big problem.
Firstly, it takes over thirty minutes to play so it isn’t the game for
starting five minutes before work! Also, it takes over half a minute to
start playing again after the end of one game (even more if you don’t
know the correct <space> sequences). The one major problem with the game
is that it is too easy. To complete all the levels it only took us a
total of six hours playing.
6.3
Conclusion
6.3
The graphics and music are excellent and you can even get the tunes to
play in the desktop. It is very addictive and the concept is good.
Although it is rather too easy, it is reasonably good value for money.
We would certainly recommend it to younger people but even older people
should not brush it aside. It is probably the second best platforms game
for the Archimedes after Chuck Rock. There have been rumours about Mad
Professor Mariarti 2 coming out and, if it does, we won’t hesitate to
buy it. Mad Professor Mariarti costs £19 from Archive. A
6.3
Preparing Material for Archive
6.3
Paul Beverley
6.3
In order to speed up the preparation of the magazine, it would be
helpful if contributors could try to follow a set of guidelines − what
you might call the “house style” of Archive.
6.3
Disc format
6.3
The text you send in can be in any disc format you like: 3½“ or 5¼”, E,
D or L format − even MSDOS if you really have to!
6.3
Wordprocessor / DTP format
6.3
We can cope with almost any WP/DTP format but, for preference, we would
like Impression since that is the application we actually use when
producing the magazine. If you do have Impression or Impression Junior,
let us know and we will send you a sample document with the styles on
it.
6.3
House style
6.3
For those of you who produce a lot of printed material, I know it’s
difficult to change your style but, if at all possible, I would like
people to prepare their articles for Archive with a particular “house
style”. I’m not talking about your style of writing − your turn of
phrase, etc − I mean the way you lay it out. A few examples will show
you what I mean.
6.3
Headings
6.3
If you look at the title of this article, you will see it is in titles,
i.e. initial capital letters for the main words. The section headings,
however, only use an initial capital letter on the first word − except
where the word would have a capital anyway, like “Impression” or
“Archive”, e.g. “How to use Impression”.
6.3
Indents
6.3
There is no need to create indents, either with spaces or tabs. This is
dealt with automatically by the “styles” used in Impression. Thus, if
you have a couple of lines of program to insert in the text, as for
example:
6.3
10 REM> WonderProg
6.3
100 PRINT “This is a load of rubbish”
6.3
110 GOTO 100
6.3
What you should send in as text is just:
6.3
10 REM> WonderProg
6.3
100 PRINT “This is a load of rubbish”
6.3
110 GOTO 100
6.3
I then give it styles which set the typeface and add the indent. If you
had already given it indents by adding your own spaces, I would have to
strip them out otherwise I would get a double indent.
6.3
Tables
6.3
If you have tables within the text and you lay out the data in column by
using spaces, it can cause problems. Remember that what looks OK in
mono-spaced type looks funny when set in a proportionally-spaced
typeface.
6.3
Thus...
6.3
Brown 6.5 13.6 11.11
6.3
Alliss 2.3 9.6 88.88
6.3
Mummy 1.1 11.1 99.99
6.3
may look OK when separated by spaces but if you put that into propor
tionally spaced text, you get:
6.3
Brown 6.5 13.6 11.11
6.3
Alliss 2.3 9.6 88.88
6.3
Mummy 1.1 11.1 99.99
6.3
If you want to use spaces to print it out, that’s fine − I then just use
search & replace to change multiple spaces into tabs. However, if you
can present it using tabs, it makes my life easier.
6.3
If you are using Impression and set up a table using a ruler, remember
that the text has to fit, if possible, into a normal column width. Also,
please don’t leave the names of any new rulers as ‘Ruler1’, ‘Ruler2’,
etc. Use your own name or some code word, like ‘Lynch1’, ‘Lynch2’, etc.
The reason for this is that when you paste text from one document (your
article) into another (the magazine) if a style of the same name exists,
it maintains the definition specified in the destination document, so
you will lose your carefully set out tabulations.
6.3
Abbreviations
6.3
I try, as far as I can, to use standard abbreviations and I try to be
consistent. Here are some I use:
6.3
a.m. (with full stops)
6.3
Basic (not BASIC − I’ve changed my mind on this one!)
6.3
e.g. (with full stops)
6.3
i.e. (with full stops)
6.3
Kbytes or just Kb (not K and not k or kbytes. Yes, that’s a change to
make it consistent with Mbytes.)
6.3
Mbytes or just Mb (and not M)
6.3
p.m. (with full stops)
6.3
RISC-OS (not RISC OS, Risc OS, RISCOS, etc)
6.3
Please refer to the computer we all know and love as an Archimedes, not
an Arc or an Archie.
6.3
Archive references
6.3
When referring to articles in previous issues, the convention is to use,
e.g. Archive 3.4 p45.
6.3
Spell-checking
6.3
If possible, please run a spell-checker over your text before sending it
in. (This may seem obvious but you would be amazed at the number of
people, even those using Impression, who don’t bother.)
6.3
Other spelling conventions that I like to stick to are: program (unless
it’s a TV programme), iconbar, filetype and drawfile.
6.3
Dashes and hyphens
6.3
A hyphen is the character on the keyboard between the zero and the
equals sign and is the thing used in hyphenated words − e.g. RISC-OS −
whereas dashes are produced as <alt-153> or from the !Chars application
and are used for separating bits of text − as used earlier in this
sentence. On !Chars, (in Trinity, anyway) it’s the third of the dash-
like characters − under the letter y. If you are using system font then
it’s the underlined d, again under the y. In Impression, you can use
<ctrl-shift-hyphen>.
6.3
To save all this hassle, all you need to do is to use a double hyphen
where you want a dash.
6.3
Describing keyboard actions
6.3
The conventions we use in Archive magazine to represent keyboard and
mouse actions are best illustrated by a bit of meaningless sample text:
6.3
If you want to press the return key (no angle brackets on return) or one
of the other keys I would tell you to press <return>, or to press <N>,
<Y> or <?>. To start up, press <shift-break> and use <ctrl-shift-f5>
(use f5, not F5) or double-click on the icon (<select> is assumed if you
don’t mention which mouse button) but click <adjust> (not “the right
hand button”) if you want to add something to the selection. In
PipeDream we have control sequences like <ctrl-BSE> for searching which
actually means pressing <ctrl-B> and then <S> and then <E> while holding
down <ctrl>. For convenience, we put them all together in one set of
angle brackets.
6.3
Comments on English style
6.3
One of the common errors that I have to correct is that you should never
use a preposition to end a sentence with. And you should never start a
sentence with a conjunction like “and” or “but”. But people do! However,
it is OK to use “however” to start a sentence. Also, you can use “also”
to start a sentence.
6.3
As a general rule, writers tend to make their paragraphs too long. It
makes the text easier to read if it is broken up into smaller logical
units. Also, it is good to use (short) titles at regular intervals to
make the structure of the article easier to appreciate. This is helpful
for busy people who haven’t time to read every word of every article and
just want to find the bit of a review that declares itself to be the
“Conclusion”.
6.3
Writing reviews
6.3
If you are writing a review, it is important to think yourself into the
position of the person reading the review who has never heard of the
product. If they are thinking of buying a copy, they need to know, at
the very least, what the item is, how much it costs and who the supplier
is. It is also useful if you can mentioned the version number of the
software under test so that people can see whether it is the current
version that was being tested or an earlier version. A
6.3
Turbo Drivers and the BJ10ex
6.3
Alan Murdey
6.3
When I saw the initial advertising campaign, back in June, I realised
that this is what I was looking for. I have had my Archimedes computer
for over a year and had managed to carry on RISC-OS printing with my old
Epson FX80, but the time had come to upgrade my printer. I had experi
ence of using LaserDirect HiRes at work but knew that I could not
justify the cost of a laser printer for use at home.
6.3
I had heard about the Canon bubble jet printers − the BJ10ex won “Best
Printer” in the Computer Shopper awards, and the TurboDrivers sounded
like a useful tool to speed up printing. Unfortunately, the product was
not available at the time of the first adverts − Computer Concepts were
apparently having problems with the software, and the release date was
put back several times.
6.3
I received a demo page from Computer Concepts, which was simply stunning
− it really was as good as laser quality. To convince myself, I had to
wet my finger and drag it across a corner of the page to prove that it
was printed with ink and not laser toner!
6.3
Finally, I brought my BJ10ex with cut-sheet feeder and TurboDriver. The
BJ10ex and cut sheet feeder came with their normal multi-language
instruction books, and the printer was soon set up, and ready to go.
6.3
Now to the TurboDriver. In the box comes the TurboDriver software
(V2.07) on one disk, an A5 instruction manual, a !Fonts disc containing
Corpus, Trinity and Homerton font families, and a 1.8m printer cable.
6.3
Installation of the software is simple. It is recommended that the
software is transferred and run from a hard disc. Once !BJ10e is on the
iconbar, printing is easy and quick. A print queue is held in the !BJ10e
application directory on the hard disc, so the once the program has
created a printfile in the queue, the printing continues in the
background freeing up control of the machine to allow further work to
continue. Clicking <select> over the !BJ10e icon produces the status
window:
6.3
6.3
Once desktop control has been handed back, this window will display the
size of the buffer. Clicking on Flush will stop printing, clear the
buffer and eject the paper being printed. A second click stops the paper
being ejected.
6.3
Background printing can be performed by using the TurboDriver to create
a Printfile which can then be dragged onto the printer icon. I think
that printing directly is just as quick.
6.3
!BJ10e will look very familiar to anyone who has used LaserDirect. The
dialogue boxes are very similar and offer the same facilities.
6.3
Useful features, again from LaserDirect, include a warning if the
printer is unable to receive data, because it could be off line or out
of paper. In RISC-OS3, you will be informed if the printer is switched
off or out of paper.
6.3
Text files can be printed by dragging them onto the printer icon. These
files are printed out in the font chosen in the Preferences dialogue box
which is called up by clicking <select> over the printer icon.
6.3
Preferences can be changed just for one session or saved permanently as
default settings for the next time !BJ10e is loaded.
6.3
I have not done any comprehensive tests but I printed a simple business
letter on my ARM2 machine. It was a PipeDream file with a graphic file
as a header, and it took only 30 seconds before control of the desktop
was handed back. Then, having created a queue of 142Kb, it took a
further 120 seconds to complete printing.
6.3
TurboDrivers will not work with the Star SJ48. Although it appears
similar to the Canon BJ10, it does not have the compatibility with Canon
BJ130e, which TurboDrivers use. Computer Concepts have included a
hardware protection, as the TurboDriver software will only work with the
special cable provided, due to the way that the data is passed to the
printer.
6.3
A problem with bubble jet printing is that the paper is sprayed with
tiny droplets of ink that must dry. My experience is that the ink dries
very quickly with text, but blocks of blacks and greys in graphics take
some time to dry. It may be possible to see that the paper is wet when
it emerges from the printer. Once the ink has dried, it can only be
smudged by a wet finger, not just by normal handling, e.g. folding over
a letter before placing it in an envelope.
6.3
Choice of paper is important because some papers will deal with the ink
better than others. So far, I have found that reasonable quality 80gsm
copier paper gives good results. Some papers allow the ink to bleed into
adjacent fibres, giving the printed page an untidy appearance. Other
people I have spoken to suggest the cheapest copier paper is best.
6.3
The Driver will not print rotated text under RISC-OS3, so some users
will be disappointed by this. However, I gather that this can be worked
around by use of FontDraw, supplied with Impression or DrawPlus.
Hopefully, Computer Concepts are working to produce a fully RISC-OS3
compatible driver. Also, they have yet to complete the RAM printing
option, where the print queue is held in RAM. This is only recommended
for users with 4Mb of RAM, with at least 1024Kb of free memory. This
option can be used now by changing the default directory for the Print
Queue to the RAM filing system.
6.3
It is possible to spot the difference between the output of the
bubblejet printer and a laser printer, but if you want laser quality at
less than a third of the price of laser printer, then have a look at
TurboDrivers and the BJ10ex. Also they could make a perfect portable
printing option, if you have an Acorn A4 − all you need is a battery-
pack costing £46 through Archive.
6.3
In conclusion, I think that I have moved from the slow lane to the fast
lane of printing using my BJ10ex and TurboDrivers. I no longer have time
to go away and make a cuppa, while my old dot matrix printer slowly
churns away at RISC-OS printing. Congratulations to Computer Concepts
for another excellent product. A
6.3
PipeLine
6.3
Gerald Fitton
6.3
Next year Easter Sunday will fall on 11th April. In the year 2000, it
will be on the 23rd April and the 16th April 1066 was an Easter Sunday!
How do I know − I have a custom function which tells me so.
6.3
I want to spend some time this month making a few comments about using
RISC-OS 3 printing from PipeDream (hereinafter called ‘Pd’ because the
abbreviation ‘PD’ has other connotations. Ed) but first a few words
about Wordz, Colton Software’s new word processor, and then something
about Easter.
6.3
Wordz
6.3
Colton Software intend to release Wordz early in 1993 (probably in
January 1993) at a price of “less than £100”. I have been assured that
registered users of Pd (3 or 4) will be offered a discount on the list
price so, if you think that you just might buy it, make sure you are
registered with Colton Software. You don’t need a registration form to
register. Just send them a note quoting your name, address, registration
number and, if you can remember it, where you bought it! At the same
time, ask to be put on their list for the “Special launch offer” of
Wordz. If you are a subscriber to PipeLine then mention that too. There
is no obligation to buy, so you might as well get your name on the list.
6.3
Published features
6.3
The handout I have from Colton Software includes the following comments.
Wordz contains all the usual features of a word processor such as
WYSIWYG outline fonts, cut and paste, search and replace and user
dictionaries. In addition, it contains many Impression-like features
such as styles and effects, rulers, a variety of tabs (left, right and
centre align) and margins and multi-column layout but without the frames
of Impression. Wordz will accept files in Pd format (and many other
formats).
6.3
Wordz will accept graphics files. These are embedded in the text and so
move with it as the file is edited. Tables can be constructed in Wordz
or imported from Pd. Different styles (including column boundaries) can
be applied to different parts of the same table.
6.3
I quote: “Wordz is the first of a family of applications . . . . all
family members have live data linking”. My interpretation of this
statement is that there is more to come − but I wonder what?
6.3
Wordz does not come with any fonts because it is assumed that you will
have outline fonts either from another application or bought as a
starter pack. Note that a font manager and a starter set of outline
fonts, printer drivers, etc are supplied with RISC-OS 3.
6.3
The handout says that background RISC-OS printing will be included. If
it is, this will save a lot of time.
6.3
My experience
6.3
I have what is described as an “alpha release” version of Wordz. As a
word processor, it is nearer to Impression than Pd. It is totally
WYSIWYG with regard to text sizes, page layout, graphics, etc. I am
pleased to be able to dispense with the frames of Impression, place text
and graphics anywhere and drag the borders of the object around
afterwards.
6.3
For example, when I want to change the margin of a block of text, I just
mark the block (by dragging) and then drag the margin. The block of text
reformats immediately. If I want to change the font in any way such as
Homerton to Trinity, or change to bold or italic, then I just mark the
block and select the font (or attribute) from a menu.
6.3
The size of the paper (where the edges are) is shown on screen. One of
the reasons I often used Impression rather than Pd was because, with Pd,
I wasn’t sure where the edges of the paper were.
6.3
One of the things I’ve not liked about Impression for letters is that
the graphic of my letter heading is saved in every Impression document.
Impression documents use up a lot of disc space. In the past, for
letters, I’ve used Pd and put my letter heading (12Kb) in my
!PipeDream.Library directory. The main reason is that I can get more
letters on a floppy disc that way. I’ve already started using Wordz
because one copy of an often used graphic can be kept in a Library
(defined as being on a Path) or on my Letters floppy and used for all
letters, fliers and mail shots.
6.3
When it comes to defining styles or adding effects, Wordz wins hands
down over Impression for speed and flexibility. Mind you, the reason I
like it is that, in Wordz, it’s all done with the mouse on the screen −
you see it happening as you do it rather than using the textual
interface of Impression. I know (because you’ve told me so) that some of
you prefer not to use the mouse. If so, you may prefer defining styles
in Impression.
6.3
Laying out tabulated data in Wordz is much easier than having to play
with rulers (and styles) in Impression. If you switch on the Wordz grid
(you can switch it on just for the part of the tabulated data you are
reorganising) you can drag the borders of individual cells left and
right or up and down until the data fits the space available in the best
possible way. In Impression, I would have had to define a multitude of
different styles (or rulers) and there’s always the problem that
something somewhere else might move when I didn’t want it to. With Wordz
I can happily lose track of how many different rulers I have had to use
to ‘fit’ a difficult table into a tight space.
6.3
The future
6.3
Information about Wordz is scarce, so I can’t be sure exactly what it
will do when it’s released. However, from what I’ve seen and from my use
of the “alpha release” version, I like it better than Impression...
6.3
(a) because it gets away from frames − it uses the whole page
6.3
(b) because the effects are much easier to apply − with the mouse
6.3
(c) because it doesn’t gobble up disc space
6.3
(d) because I can see what’s happening on screen as I use the mouse
6.3
I am sure that spreadsheets and charts created in Pd and loaded into
Wordz will be hot linked. This means that I will be able to use Pd files
(including custom functions, charts, etc) as dependent documents of a
Wordz document. Having created the structure of my application (i.e. a
Wordz document supported by Pd tables and charts), I will be able to
print from the Wordz document with the assurance that comes from seeing
the layout of the page (including the borders) on the screen while still
being able to modify it through the hot linked Pd spreadsheets and
charts.
6.3
One problem that I have with software is coming to terms with the idea
that my favourite package begins to look dated just as soon as I begin
to feel that I’ve mastered it. Do I stick with Pd, Impression and
DrawPlus or do I get something new and have to go through the painful
and time consuming process of learning how to use it?
6.3
I’m glad I changed from Wordwise to Pd and from Acorn’s DTP to Impres
sion. If Wordz fulfils its promise, I think I’m going to change from
using Impression to using Wordz as the document from which I print − but
it’s too soon for me to be sure yet. If I do decide to go for Wordz, one
thing which I must be able to do is to port across my old Impression
documents into Wordz without having to do too much work on them
afterwards.
6.3
Let me know if you have any specific questions you want answered about
Wordz, particularly if it is to do with the Pd − Wordz link and I’ll see
what I can do.
6.3
Easter
6.3
Writing custom functions is fun! Don’t get me wrong, I know that word
processing is the most ‘popular’ use of modern PC systems (and, having
used one, I wouldn’t like to be without it) but, let’s face it, WP
systems need a lot of data input, do relatively little ‘processing’ and
produce a fairly predictable output. Now, with a ‘Day of the Week’
program you can find out if you are “Full of Grace” (Tuesday’s child) or
whether you have “Far to Go” (Thursday’s child). Much more exciting!
Why? I believe that custom functions are more exiting to use than WP
packages because the ratio of processing (by the system) to data input
(by you) is much higher and because the output is less predictable!
6.3
Finding the date of Easter Sunday has a rather more intricate algorithm;
I believe it is the first Sunday after the first full moon after the
Spring Equinox and it ultimately bears some relationship (via the Jewish
feast of the Passover) to the date of the tenth Plague of Egypt (see
Exodus for further details). Traditionally (manually) the date of Easter
is calculated using, as intermediate variables a Golden Number, a Sunday
Letter and an Epoch, all with appropriate look-up tables.
6.3
In older records of the birth, marriage or death of individuals or in
the records of important events, the dates are often related to the
nearest ‘movable feast’; the dates of all these ‘movable feasts’ can be
found once you know the date of Easter. For example, in the year that
Columbus sailed off to America, 1492, Easter Sunday was on the
22nd April. If you are a genealogist or historian, knowing the date of
Easter can be important. If it is important to you (or someone you know)
then using a custom function to find the date of Easter can be fun! Why?
Because, generally, it would take too long to calculate the result
manually and you would be left wondering if the result was reliable.
6.3
On the Archive monthly disc is my attempt to write a custom function for
finding Easter Sunday for any year from about 5AD onwards. I have had
helpful advice from over a dozen people and I expect more advice to
come. However, I must single out for special mention Brian Edwards,
Nigel Kinsella and Danny Fagandini. The algorithm is not as elegant as I
would like − but it works. As written, it assumes the Julian calendar
until 1582 and the Gregorian calendar thereafter (but the year can be
changed by amending one line of the custom function). In the UK, we
changed from Julian to Gregorian in 1752 by leaving out 11 days between
the 2nd and 14th September (i.e. the 2nd September 1752 was followed by
the 14th September 1752) so, between 1582 and 1752 my algorithm doesn’t
apply to the UK! If you have any comment or contribution to make as to
whether Easter was celebrated simultaneously on the continent and in the
UK between 1582 and 1752, I shall be most pleased to hear from you.
6.3
I had thought of including the custom function as text this month but I
have decided to wait until those of you who take the monthly Archive
disc have written back to me with improvements!
6.3
RISC-OS 3 printing
6.3
Now, let me turn to RISC-OS printing. Pd uses RISC-OS printer drivers in
a slightly different way from many other packages. For example, if you
RISC-OS print a Landscape page from Paint, Draw or Impression, all
calculations about where to print are based on an origin at the bottom
left of the ‘page’. Pd uses the top left of the page as its origin.
6.3
Of course, provided that you have the ‘correct’ page length and margins,
it doesn’t matter whether you use the bottom left or the top left as
your origin but, if you have either the margin or the page length
slightly wrong, you might be hard pressed to discover how to put it
right! On this subject, many page printers ‘stretch’ the paper slightly
as it passes through the rollers so that the paper is about 0.5% larger
than your printer thinks it is!
6.3
My advice about setting up your printer driver is to do so in Draw
rather than in Pd since, if you get that right, you will probably be OK
for Impression, Wordz and Pd. On the Archive monthly disc is a file in
Draw format called [Rectangle] to print out from Draw.
6.3
You will need to copy onto your ‘Printers’ disc (or your hard disc if
that is what you use) a directory called !Printers, another called
!PrintEdit and a printer definition file for your printer − mine is
called [LasJet-II] so, where I write ‘LasJet-II’ you will have to
substitute your own printer definition. Make sure that RISC-OS is
recognising these copies of your applications (and not the originals) by
double clicking on !Printers and on !PrintEdit.
6.3
If the version of !Printers which appears on the iconbar says LasJet-II
(instead of !Printers), click <menu> over the icon and click <select> on
printer control. When the printer control menu appears, click <menu> on
it and remove your printer definition.
6.3
Double click on the file [LasJet-II] and a printer definition editor
window will open. The fifth row is the paper x-offset and the sixth row
is the paper y-offset. You may have to fine tune these as I describe
below but, for now, just a word of warning; strange things happen if you
make them zero or very much different from the paper size graphic
margins (see below). To save an amended printer definition, click <menu>
over the printer definition editor window and run the pointer through
save followed by clicking <select> on the OK box.
6.3
Having saved an amended printer definition, you have to load it into the
printer control window (see above) by dragging the [LasJet-II] file into
the printer control window. Note that your printer will not use the new
paper x-offset and paper y-offset unless you remove the printer
definition from the printer control window and load (by dragging) the
amended version saved from the printer definition editor window! I don’t
think it is easy to get the sequence of ‘actions’ right first time and
there is little indication of the values of the offsets you are using!
6.3
The paper size window is accessed by clicking <menu> over the installed
(on the iconbar) !Printers icon and clicking on Edit paper sizes. When
the paper size window opens, click on the right arrow just to the right
of the dialogue box at the end of the first row (called paper size
name). Select a suitable paper size such as A4 (Generic DP). You can set
the top, bottom, left and right graphics margins from the paper size
menu. Remember to click on the OK box.
6.3
Your amended graphics margins will not be saved to disc until you save
choices from the !Printers icon (use <menu> over the !Printers icon to
find Save choices). The default paper size and its graphics margins are
completely independent of the paper x-offset and paper y-offset referred
to above! The paper size and graphics margins apply to all printer
definitions that you may load (by dragging) into the printer control
window.
6.3
Another word of warning. Do not set the graphics margins to zero or
unpredictable things will happen!
6.3
What I have found in Draw is that as I increase the graphics top margin
my [Rectangle] file is printed lower down the page. As I increase my
paper y-offset the [Rectangle] file is printed further up the page. The
‘correct’ setting is such that the [Rectangle] is printed on the
(physical) page at the coordinates shown on the screen in Draw and that
the graphics top margin and the paper y-offset are about equal and both
(approximately) equal to the physical ‘margin’ at the top of the paper
where the printer will not print.
6.3
One problem I had initially is that, if the paper y-offset is larger
than the graphics top margin, the ‘extra’ push upwards that I wanted
didn’t happen. I discovered that by slightly shortening the page (from
297 mm to 295 mm) in the paper size window and reducing the graphics
bottom margin to compensate, I could achieve a slight upward reposition
ing of the printed [Rectangle]. This artificial shortening of the
(logical) page also stopped my printer feeding out an unwanted page from
Impression.
6.3
I was fortunate with my graphics left and right margins in that I set
both of them and my paper x-offset (all three) to 5 mm and my
[Rectangle] file printed on the paper exactly where it appeared on the
Draw screen, so I am not sure about the inter-relationship between these
three parameters. However, there is no doubt from a quick test that both
the graphics left margin and the paper x-offset have independent
effects!
6.3
Now to Pd. My tests show that if you get the [Rectangle] file printing
correctly in Draw, Pd will print out (using the RISC-OS 3 drivers) where
you think it should! By this I mean that if you use a zero Pd top margin
and a zero Pd left margin (use the <Ctrl PY> command to open the window
for making these settings) then a graphic which is placed in slot A1
will print with its left top point at the paper x-offset and paper y-
offset position. I think that Pd does not make use of the paper size
graphic margin settings but only the paper offset values.
6.3
In conclusion
6.3
I don’t know about you but I found setting up the RISC-OS 3 printer
definitions and paper size very difficult. I have received more requests
for help on this topic than I have received for a long time! I’m still
not sure that I’m doing everything the ‘right’ way. I keep thinking that
it ought to be a lot simpler than it is. If you have anything to
contribute to this discussion, I shall be most grateful if you will
write to me at the Abacus Training address which is on the inside back
cover of Archive. A
6.3
Bestform − DTP Formulae Editor
6.3
Lothar Böhler
6.3
Bestform is a fully RISC-OS compliant editor for mathematical formulae
which costs £49 (+VAT) from Klein-Computer in Ruesselsheim, Germany.
6.3
How to use it
6.3
Like every RISC-OS application, Bestform is started by double-clicking
on the icon in the filer window. It then puts an icon on the iconbar
which, when clicked, opens one (or several) windows. The window shows
the usual red caret at which your typed text will appear. When the text
becomes too large to fit in the window, it will be enlarged (quite
usual). By clicking <select>, you can position the caret, by clicking or
dragging <adjust>, or dragging <select>, you can mark a block − still
usual, except that sometimes some experience is needed to select exactly
the desired block. Furthermore, you can select a block, i.e. a sub-
formula, by double- or triple-clicking on one of its components. As may
be familiar from Edit and the new Draw of RISC-OS 3, the marked block
can be deleted (with <ctrl-X> or by choosing the related Menu entry), or
copied (<ctrl-C>) or moved (<ctrl-V>) to the caret position. You can
even copy or move between different windows, save the selected block or
load a file at the caret position. This is a great advantage over most
other programs that use their own system of move/copy/delete − one has
to learn anew − and many are not as elegant or powerful.
6.3
There is another unusual thing. With each of function keys F1 to F4, one
of four different fonts can be selected. As default these are Trinity
and Mathgreek, each in plain or italics. You can set it up to choose
four different fonts, if you wish.
6.3
Creating formulae
6.3
The most important parts of a formula are the easiest to create in
Bestform − indices and exponents. Just press cursor-down and the caret
is in the index-position or cursor-up for exponents. With page-down/up
you can toggle between these positions and even ‘walk’ around the
variable or sub-formula to enter a front index or a front exponent. To
get the caret back to the original position, just press cursor-right.
6.3
Other structures are available from several pop-up windows, which are
either opened by choosing the relevant menu entry or by pressing the
correct function key. You can either first choose the structure and then
fill in the components or select a block and then choose a structure. In
the latter case, the selected block will be used as one component. Any
structure can be removed anywhere by selecting the smallest block that
contains the structure and choosing ‘DEL’ from the relevant pop-up
window.
6.3
What structures are available?
6.3
There are a lot of structures available, arranged in four groups:
brackets, big operators, structures and attributes. Each group has a
pop-up window which can be opened with a function key (F5 − F8). There
are ten brackets and a special type which allows combinations of the
other ten brackets for the left and the right side. (See example below.)
By choosing no bracket for one side, you can get single brackets.
6.3
Among the big operators are: sum, integral, circle integral and product
operator − each available either with limits above and below or limits
as subscript/superscript. Also, there are root and/or union, intersec
tion and other big operators − a total of seventeen.
6.3
The pop-up window “structures” contains three types of fraction (normal,
double, slanting), over- and underwrite, over- and underbrace and
matrices (!). Over- and underwrite are very useful e.g. in equations,
you may overwrite the equal sign with a short comment. Parts of a larger
formula can be commented very pleasantly by using the overbrace
facility. The matrices are perhaps the most powerful feature of Bestform
(and the most painfully lacking facility in Equasor). Conceptually, a
matrix in Bestform is just a rectangular structure of up to 9 by 9 sub-
formulae. You can use matrices for nearly everything: normal matrices
(enclosing them in brackets), determinants, vectors (when using 1-by-n
matrices), equation systems (using 3-by-n matrices), case distinctions,
etc. There is a built-in matrix editor for deleting, inserting, copying,
moving and formatting rows and columns.
6.3
As ‘attributes’ there are: back and forward arrow, tilde, point, double
point, accent, and under- and overline. All attributes can span a whole
sub-formula. Except for the point attributes, they are all enlarged
automatically to the length of the sub-formula.
6.3
What about additional characters?
6.3
There are 36 arrows, 23 comparisons, 17 sets, 19 operators and 26
special signs. These are not part of an outline font but are defined as
draw objects (as well as the big operators and structures already
mentioned). This has the disadvantage that, on the screen, they are not
plotted with anti-aliasing. However, there are several advantages: they
do not need any font cache; you can easily add your own signs by
defining them with Draw; and they are drawn much faster because it is
not necessary to calculate a bitmap. This is especially noticeable when
a sign or operator is needed in a new size (e.g. because you have
entered a big operator in the index). With Equasor, it is very likely
that you will see the hourglass for several seconds, but this never
happens with Bestform.
6.3
Finally, there is another window for entering Greek letters. (A Greek
outline font is included with Bestform.)
6.3
Additional features
6.3
Version 1.1, which I possess, supports printing. Unfortunately, it does
not work with the new RISC-OS 3 printer drivers. An attempt to print
caused a total system crash (error message: “font handle out of range”)
which was not even recoverable with Quitter (a utility which kills the
current task when it hangs) and destroyed my first version of this text
− what a shame! (So you can see how much this program impressed me, as I
wrote this review all over again). With the old printer driver, printing
seems to work perfectly well.
6.3
In a special window, you can choose the four fonts which are accessible
through F1−F4 and the size of the normal text, the index text and the
second index text. The third and further indices have the same size as
the second.
6.3
Bestform has no special filetype. It always saves its files as draw
files. To be able to reload and reconstruct the formula from the
drawfile, it saves it together with so-called tags. So you can save the
formula directly (RAM-to-RAM) to other programs, e.g. Draw or Impression
and back again! Unfortunately, some programs do not cope with tags
(shame on them − they are defined and explicitly allowed by Acorn) so
Bestform also supports saving without tags.
6.3
Comparison with Equasor
6.3
Bestform is a very powerful formulae editor with all the features one
needs (I guess) when editing a formula. I think this program is a reason
to forget TeX and Equasor. It is much easier to use than the former and
much more powerful than the latter. As Bestform is quite similar to
Equasor, I will compare these two but note that I only possess Bestform
and not Equasor, which I only know from some demonstrations. So, if I
should make incorrect comparisons, please do not blame me too much. (I
have asked Brian Cowan to comment from his knowledge of Equasor − see
below. Ed.)
6.3
Of course, not everything is perfect with Bestform. For example, I am
not quite sure how systems of equations should be entered. You must
create a matrix, which has then to be formatted: left column right
justified, mid column centered, right column left justified. This is a
bit annoying but you can create an empty matrix with the correct
formation, save it and use it again and again. One possible problem is
that it cannot create matrices larger than 9 by 9. If you need larger
equation systems, you must either create a matrix of matrices, or split
the system into several parts which are then rearranged in Draw or
whatever DTP program you use. So, the method used by Equasor seems to be
more straight-forward but, on the other hand, with Bestform’s method you
can be sure that your equation system is always perfectly formatted and
that rows do not overlap − no matter what heights the different rows
have.
6.3
I was really disappointed with Equasor as far as speed is concerned.
Entering a sign or operator in a new size, deleting rows and displaying
the Greek font window usually activate the hourglass whereas, with
Bestform, all these things happen immediately. (Could the Equasor demo
have been done with too small a font cache? Brian seems to think
Bestform is slow − see below. Ed.) This is because Bestform does not use
fonts for its signs and operators but draw objects and sprites − the
latter only in the pop up windows of course, not for displaying or
printing formulae! On the other hand Bestform always redraws the whole
window when something is entered. This is not very pleasant, if you have
typed in a large equation system, because redrawing is not very fast
(unless you have an ARM 3).
6.3
Another disadvantage of this system is that the operators, signs, etc
are less pretty and easy to read because they do not gain from the anti-
aliasing of the font manager. Everybody should decide for himself how
important this is − for printing, anti-aliasing is not used anyway.
6.3
Equasor supports four or five different sizes of indices whereas
Bestform supports only three, but is it sensible to make the indices of
third and fourth degree smaller and smaller?
6.3
Bestform unfortunately crashes on some occasions. How often does Equasor
crash? I do not know because I have no real experience with it.
6.3
With Bestform, you cannot create several independent formulae in one
window and position them. Also, printing does not work with RISC-OS 3
printer drivers but these are not real disadvantages because it is most
likely that you will export the formula to a DTP program where you can
arrange and print as you wish. Here Bestform has the advantage that
there is no need to save the formula separately for further editing −
you can reload the formula as a drawfile directly from the DTP program
(if it supports this feature).
6.3
Furthermore Bestform is much easier to use, especially in the case of
creating indices or exponents.
6.3
Finally, Bestform beats Equasor in terms of the range of facilities
available: attributes spanning whole sub-formulae, under-/over-brace,
arrows which can easily be connected, vectors, matrices and case
distinctions. I have seen an advert for Bestform with at least a dozen
(sensible) formulae of which Equasor could only create one!
6.3
Conclusion
6.3
Though Bestform is not perfect (what program is?) it is, in my opinion,
the best for editing formulae (not only on the Archimedes either). It
has a lot of facilities, is fast enough for convenient working, is very
easy to use and is the same price as Equasor. A
6.3
Brian Cowan adds: I think that Lothar Böhler’s review of BestForm is
very fair. I agree with most of the points he makes. However, having
tried an earlier version of BestForm (version 1.0) I found many problems
with it and I returned to using Equasor. I will not detail these
problems as it is likely that most have been rectified in version 1.1.
Assuming this is so, then BestForm will have the edge on Equasor and I
might well be changing in the future.
6.3
While BestForm follows the conventions of Draw and Edit for dealing with
selected objects, Equasor follows the conventions of Impression. Thus
Equasor is more intuitive for Impression users. With Equasor, the
characters from other fonts are not entered from the keyboard but by
mouse with a “symbol picker” like !Chars. You can have many more than
four fonts, which is useful. This is a serious drawback of BestForm.
6.3
The treatment of subscripts and superscripts is different for BestForm
and Equasor. As Lothar explained, with BestForm, you simply move the
cursor up or down but, with Equasor, you must select a region to which
the subscript or superscript is to be applied. This is slightly more
inconvenient but the result can look better.
6.3
Equasor has its own filetype and objects can be saved either as Equasor
files or as drawfiles. The Equasor file is essentially the drawfile
together with tags (rather like BestForm files). Programs such as
Impression accept Equasor files but since, as Lothar mentions, some
programs do not like drawfiles with tags, the bare drawfile format is
then useful.
6.3
The real area where BestForm wins is in the range of structures
available. In particular, the lack of matrices in Equasor is a serious
omission. I generally produce matrices by creating parts in Equasor,
saving them as draw objects and then combining them using Draw but this
is rather tedious.
6.3
I found BestForm to be slower than Equasor in use. I was particularly
annoyed by BestForm re-drawing the window after each character is
entered. I also had BestForm crash my machine from time to time.
However, Equasor also occasionally crashes but it never hangs the
computer; it just kills itself. I have certainly found Equasor to be
much more reliable.
6.3
Equasor allows for many equations on one page. This can be useful if you
want a number of equations bracketed together on a page but it can also
be useful for another reason. In Impression, each graphic frame, such as
an equation, is stored as a separate file. However, there is a limita
tion on the number of files which may be stored in a directory. I once
lost an entire document by exceeding this quota. A solution recommended
by Computer Concepts is to have all equations in one large file and then
to open up different windows on different parts of the page. You
certainly can’t do that with BestForm.
6.3
I was hoping that a new version of Equasor would be appearing, rectify
ing the various omissions; that now seems unlikely. For anyone who
already has a DTP package and who wishes to use it for scientific or
mathematical work, I think Bestform is the natural choice. The file
number limitation can be overcome either by splitting a document into
smaller parts or by using one of the disc compression systems where all
files are stored in a single archive file; the number of files in such a
compressed “directory” has no limit. A
6.3
Princeton High Resolution Monitor
6.3
Andrew Ferguson
6.3
I have owned an A3000 and a Phillips CM8833 colour monitor for the last
three years. The CM8833 is a good monitor but it does not allow you to
take full advantage of large desktop screen modes on the Archimedes. In
order to use these screen modes, you really need to have a ‘better’
monitor.
6.3
The Archimedes can be used with four different types of monitors,
standard, multiscan, SVGA and high resolution (monochrome). The User
Guide shows a list of screen modes (Appendix D in my manual) and the
monitor types that support them. A multiscan monitor offers the largest
choice of screen modes but the same modes are available on a Super VGA
(SVGA) monitor if you use the Watford SVGA VIDC Enhancer.
6.3
I recently purchased a Princeton Ultra 1400 monitor for the sum of £245
including VAT, (I believe that this is by far the cheapest SVGA /
multiscan monitor on the market). The monitor’s specifications are as
follows:
6.3
Tube 14 inches diagonal, 0·28 dot pitch, black matrix tube
6.3
Video bandwidth 45 MHz
6.3
Resolution Horizontal: 1024 dots, Vertical: 768 (SVGA)
6.3
Sync Signals TTL positive or negative, composite or separate
6.3
Synchronization Horizontal 15 − 36.5 kHz autosynchronous
6.3
Vertical 45 − 120 Hz autosynchronous
6.3
The monitor has a 15 pin VGA connector on the rear and was supplied with
cables for an IBM PC Compatible and a Commodore Amiga. The PC cable
supplied is a 15 pin male to female VGA cable and a 9 pin to 15 pin
adapter cable. Do not fall into the trap of using the supplied PC 9 pin
cable adapter, as the PC VGA ports are not pin compatible with the
Archimedes. Having said this, the supplied 15 pin cable should work on
the A5000 and the A4, as the RGB port is the same as on a PC. Watford
Electronics can supply a 9 pin to 15 pin adapter cable for the Archi
medes for £15. (Or £12 through Archive!)
6.3
The monitor has a very sharp picture and has all the picture control
buttons on the front of the monitor. The controls allow you to change
the vertical and horizontal position / size. It also allows you to
change the brightness / contrast and text colour (which can be useful
with the PC Emulator).
6.3
Without a VIDC adaptor you can use the monitor with the built in VGA and
multiscan modes. However, for serious use you need a VIDC enhancer (I
bought the SVGA VIDC Enhancer, but the Princeton may work with an
ordinary MultiSync VIDC adapter).
6.3
A VIDC enhancer is a small card that speeds up the clock speed of the
VIDC. The SVGA card provides additional oscillators of 25.175MHz and
36MHz to give faster line rates. As the VIDC is working at a higher
clock rate, it can manage ‘larger’ screen modes i.e. Mode 102 (1152×480
at 16 colours), compared with Mode 12 (640×256). As you can see, these
screen modes are much larger and take more processor bandwidth to
refresh. In practice, this means that if you do not have an ARM3
machine, the large modes may be too slow to use.
6.3
The VIDC comes with a disc of utility programs, a small daughter board
for the computer and comprehensive instructions.
6.3
On the A3000 installation is straight forward. The case has to be opened
and the keyboard and floppy disc drive removed. The daughter board is
fitted onto links LK28 and LK29 and two probes are connected to the
external expansion connector. The only problem that I experienced was
with the power connection. You have to unplug the red power spade and
connect to the motherboard as in the diagram.
6.3
6.3
As you can see from the picture the original Power Lead connects at a
45° angle and, if like me, you have a Serial Port sampler/midi card
attached to the Econet socket you will have problems. Watford have
suggested that I bend the original connector so that it fits under the
sampler/midi card. I haven’t tried this yet as it could obviously break
the power connector.
6.3
If you do not feel confident to attempt the installation, the manual
suggests that your local Acorn dealer, or Watford Electronics, can fit
the board for you. There will be a small charge for this service.
6.3
Once it is fitted, you can turn on the machine and if you fitted
everything correctly, the desktop should appear as before. In order to
use the new modes and facilities of the VIDC Enhancer, you need to use
the supplied software.
6.3
The disk comes with 4 applications, !Define, !ModeSel, !Tweaks and
!VIDCMods.
6.3
!ModeSel − This application allows you to select any RISC-OS wimp mode
(including the new VIDC modes). To use the program you need to select a
base monitor and a large selection of common monitors are included, e.g.
Aries VGA, Eizo 9060S and 9070S, NEC 2A, 3D and 4D, Taxan 775 and 795.
The Princeton is not included and to cater for other monitors, modules
for ‘Base’ monitors are provided. These are:
6.3
BaseVGA: 640×350, 720×400, 640×480 VGA mode monitors
6.3
Basesuper: 640×350, 720×400, 640×480 & 800×600 VGA mode monitors
6.3
BaseTV: 15.625kHz line rate, 60.080Hz refresh rate TV monitors
6.3
BaseSync15: 15kHz+ line rate multisync monitors
6.3
BaseSync20: 20kHz+ line rate multisync monitors
6.3
BaseSync30: 30kHz+ line rate multisync monitors
6.3
After experimenting with various monitor modules, I selected a TX-14H8BT
module.
6.3
This gives access to about 55 screen modes which includes all suitable
Acorn RISC-OS modes, Computer Concepts !ExtraModes and the Atomwide VIDC
Enhancer modes.
6.3
!Tweaks − This is a superb utility that allows you to move the displayed
image on the monitor’s screen. One of the problems with multiSync / SVGA
monitors is that, as they lock onto a new frequency following a mode
change, the top left hand corner of the screen display moves. So
changing mode means that you can no longer see all of the screen and you
have to use the monitor controls to adjust the display. This is a
nuisance, especially if the monitor adjustment switches are on the back
of the monitor.
6.3
With !Tweaks you should never have to touch the monitor control buttons
because, every time you change mode, you can select the screen perfectly
centred. In order to do this, you use a Real Time Move option, which
allows you to move the screen display by moving the cursor keys until it
is centred. You can save the screen mode. As far as I am aware, this
facility is not available on any of the other VIDC enhancers.
6.3
!Define − If you do not like any of the 55 supplied screen modes, you
can change them or create new screen modes. You can select the number of
rows and columns, the number of colours (2, 4, 16 or 256), the clock
speed and VIDC rate. (I have not really used this utility as I am quite
happy with the existing modes.)
6.3
!VIDCMods − This is an application which is similar in nature to the
!System folder. It contains modules holding all the screen modes for the
monitors supported by the other applications.
6.3
Conclusion
6.3
I am very impressed with the Princeton monitor. For the price, it is
superb and its features are comparable with monitors costing twice as
much. The resolution is very good and it is great having large screen
modes.
6.3
However, I have just turned my ARM3 off and the computer is very
sluggish in Mode 21 (640×512 256 colours). You may want to consider
upgrading to an ARM3 before you use these screen modes. With an ARM3,
this monitor is a real bargain.
6.3
I am also impressed with the Watford SVGA VIDC Enhancer. This is worth
buying for the !Tweaks utility alone, especially if your monitor
controls are on the back on the machine.
6.3
However, there are problems if you have a Serial Port combined midi /
sampler card. I also did not like having to keep changing applications
to define screen modes and then tweak them. I would hope that, if there
is a new release of the software, then one application would be produced
which combines the three existing applications.
6.3
Supplier details
6.3
I bought the monitor for £210 + VAT from Sudbury Electronics, 771 Harrow
Road, Sudbury, Wembley, Middlesex HA0 2LW. (081−908−1888) (Fax:
081−904−8002)
6.3
Additionally, other dealers can be found by contacting Princeton Graphic
Systems, Norfolk House, Great Chesterford Court, Great Chesterford,
Saffron Walden, Essex CB10 1PF (0799− 30963). A
6.3
EFF Scientific and Mathematical Fonts
6.3
Brian Cowan
6.3
Scientific or mathematical writing requires either a dedicated scienti
fic DTP product such as TechWriter (to be reviewed in a future issue) or
a general-purpose DTP program, such as Impression, together with an
equation generation tool such as Equasor or BestForm (see the review on
page 65). However, for presentable scientific manuscripts, an essential
requirement is a good set of fonts. Both Equasor and BestForm are
supplied with the MathGreek fonts from Electronic Font Foundry, which
comprises an upright and an italic style, while Techwriter comes with
font family called MathPhys which is very similar to MathGreek and is
also produced by EFF. There are two other relevant fonts from the EFF
stable, known as MathScript and MathLogic. Examples of all these fonts
are shown in the tables.
6.3
MathGreek
6.3
There are a number of important points to appreciate about the MathGreek
fonts. There are actually two versions of this font and, unfortunately,
Equasor and the version of BestForm I have both use the earlier version.
You will see from the tables that the new MathGreek font contains an
enhanced collection of symbols and so it is rather more useful. However,
you will also observe that the position of some of the characters has
changed; their ASCII codes have been altered. This means that expres
sions composed with the older version of the font may print out as
nonsense when using the new fonts with the same name. In fact, this is
the explanation of why some of Computer Concepts’ adverts for Equasor
showed gobbledigook equations containing hearts and other strange
characters.
6.3
Another feature of MathGreek is the ten characters at the left hand edge
of the tables on the bottom two lines. These characters have negligible
widths so their effect is to over-type the character which follows it.
The usual use for such symbols is in adding accents to foreign language
text but in the scientific context, this allows one to create frequently
used objects such as ãx, Ài , âv, án, etc. The difference between the
set on the two lines is their height; the upper set are higher.
6.3
Initially, I had a big moan about the lack of one particular physical
symbol that I seem to use a lot, namely Planck’s (other) constant which
is written as an h with a slash through it. Then I discovered these
“over-typing” characters and the production of äh became trivial; no
longer did I have to use extensive kerning to combine the regular h with
the _ bar.
6.3
My only unsatisfied requirement here is for bold Greek characters.
However, it is worth pointing out that the common Trinity font has a
Greek mu, and Trinity (both Roman and italic) comes in a bold flavour,
thus providing the characters µ and µ. We shall have to wait for a full
bold Greek alphabet; when that will come, I don’t know.
6.3
MathPhys
6.3
I only came across this font when I received my review copy of Tech
Writer and, initially, I did not scrutinise it sufficiently to
appreciate that this was another product from EFF. Then I noticed the
logo and discovered some of its interesting features. To my joy,
Planck’s constant is there as ⇩, and also there are now over-typing
arrows so that vectors can be represented as år, for example. Another
major addition is the multiple integral signs † and ‡. The MathPhys font
comes in Roman and italic styles but, again, there is no bold. Signifi
cant losses are the real and imaginary symbols ½ and ¼.
6.3
MathScript and MathLogic
6.3
The MathScript “curly” characters such as H and L will be of use to
physicists. I tend to use these a lot in quantum mechanical equations.
However, the straight open characters like r will be of more use to
mathematicians. Interestingly, one mathematical symbol (which “enhanced”
Archimedes users might recognise), namely », the Hebrew letter aleph,
must be found on the MathGreek font.
6.3
The set of logic symbols is quite extensive. However, for usual
scientific applications, the range of inequalities, etc on the other
fonts is fine. All in all, it seems that almost everything you could
need is somewhere. The only thing I would like (apart from bold Greek)
which, as yet, does not exist is the facility for producing fractions
with a horizontal bar. This is useful for clarity in equations. It would
actually be quite easy to do this using one set of normal “width”
characters, for the numerator, say, and a set of “zero width” characters
for the denominators.
6.3
An important point to appreciate is that EFF have deliberately avoided
being restricted by the PostScript character sets. This has resulted in
a much more versatile collection of fonts. For dot matrix printers, and
particularly for Laser Direct printers, this is no problem at all. Of
course, with the utilities of RISC-OS 3, outline fonts can easily be
downloaded to PostScript printers.
6.3
EFF are to be congratulated on the superb range and quality of their
scientific fonts. A
6.3
! “ # $ % & ‘
( ) * + , - . /
0 1 2 3 4 5 6 7
8 9 : ; < = > ?
6.3
@ A B C D E F G
H I J K L M N O
P Q R S T U V W
X Y Z [ \ ] ^ _
6.3
` a b c d e f g
h i j k l m n o
p q r s t u v w
x y z {| } ~
6.3
€ Ŵ ŵ ◰ ☓ Ŷ ŷ ⇦ ⇨ ⇩ ⇧ … ™ ‰
• ‘ ’ ‹ › “ ” „
– — − Œ œ † ‡ fi
fl
6.3
¡ ¢ £ ¤ ¥ ¦ §
¨ © ª « ¬ ® ¯
° ± ² ³ ´ µ ¶ ·
¸ ¹ º » ¼ ½ ¾ ¿
6.3
À Á Â Ã Ä Å Æ Ç
È É Ê Ë Ì Í Î Ï
Ð Ñ Ò Ó Ô Õ Ö ×
Ø Ù Ú Û Ü Ý Þ ß
6.3
à á â ã ä å æ ç
è é ê ë ì í î ï
ð ñ ò ó ô õ ö ÷
ø ù ú û ü ý þ ÿ
6.3
! “ # $ % & ‘
( ) * + , - . /
0 1 2 3 4 5 6 7
8 9 : ; < = > ?
6.3
@ A B C D E F G
H I J K L M N O
P Q R S T U V W
X Y Z [ \ ] ^ _
6.3
` a b c d e f g
h i j k l m n o
p q r s t u v w
x y z {| } ~
6.3
€ Ŵ ŵ ◰ ☓ Ŷ ŷ ⇦ ⇨ ⇩ ⇧ … ™ ‰
• ‘ ’ ‹ › “ ” „
– — − Œ œ † ‡ fi
fl
6.3
¡ ¢ £ ¤ ¥ ¦ §
¨ © ª « ¬ ® ¯
° ± ² ³ ´ µ ¶ ·
¸ ¹ º » ¼ ½ ¾ ¿
6.3
À Á Â Ã Ä Å Æ Ç
È É Ê Ë Ì Í Î Ï
Ð Ñ Ò Ó Ô Õ Ö ×
Ø Ù Ú Û Ü Ý Þ ß
6.3
à á â ã ä å æ ç
è é ê ë ì í î ï
ð ñ ò ó ô õ ö ÷
ø ù ú û ü ý þ ÿ
6.3
! “ # $ % & ‘
( ) * + , - . /
0 1 2 3 4 5 6 7
8 9 : ; < = > ?
6.3
@ A B C D E F G
H I J K L M N O
P Q R S T U V W
X Y Z [ \ ] ^ _
6.3
` a b c d e f g
h i j k l m n o
p q r s t u v w
x y z {| } ~
6.3
€ Ŵ ŵ ◰ ☓ Ŷ ŷ ⇦ ⇨ ⇩ ⇧ … ™ ‰
• ‘ ’ ‹ › “ ” „
– — − Œ œ † ‡ fi
fl
6.3
¡ ¢ £ ¤ ¥ ¦ §
¨ © ª « ¬ ® ¯
° ± ² ³
6.3
! “ # $ % & ‘
( ) * + , - . /
0 1 2 3 4 5 6 7
8 9 : ; < = > ?
6.3
@ A B C D E F G
H I J K L M N O
P Q R S T U V W
X Y Z [ \ ] ^ _
6.3
` a b c d e f g
h i j k l m n o
p q r s t u v w
x y z {| } ~
6.3
€ Ŵ ŵ ◰ ☓ Ŷ ŷ ⇦ ⇨ ⇩ ⇧ … ™ ‰
• ‘ ’ ‹ › “ ” „
– — − Œ œ † ‡ fi
fl
6.3
! “ # $ % & ‘
( ) * + , - . /
0 1 2 3 4 5 6 7
8 9 : ; < = > ?
6.3
@ A B C D E F G
H I J K L M N O
P Q R S T U V W
X Y Z [ \ ] ^ _
6.3
` a b c d e f g
h i j k l m n o
p q r s t u v w
x y z {| } ~
6.3
€ Ŵ ŵ ◰ ☓ Ŷ ŷ ⇦ ⇨ ⇩ ⇧ … ™ ‰
• ‘ ’ ‹ › “ ” „
– — − Œ œ † ‡ fi
fl
6.3
¡ ¢ £ ¤ ¥ ¦ §
¨ © ª « ¬ ® ¯
° ± ² ³ ´ µ ¶ ·
¸ ¹ º » ¼ ½ ¾ ¿
6.3
À Á Â Ã Ä Å Æ Ç
È É Ê Ë Ì Í Î Ï
Ð Ñ Ò Ó Ô Õ Ö ×
Ø Ù Ú Û Ü Ý Þ ß
6.3
à á â ã ä å æ ç
è é ê ë ì í î ï
ð ñ ò ó ô õ ö ÷
ø ù ú û ü ý þ ÿ
6.3
MathGreek (old version)
6.3
MathGreek (new version)
6.3
MathPhys
6.3
MathScript
6.3
MathLogic
6.3
SFXM − Sound Effects Maker
6.3
Peter Smith
6.3
!SFXM is the fully RISC-OS compliant Sound Effects Maker from Cambridge
International Software. It actually comprises two applications − the
Sound Effects Maker itself and a waveform editor, !SFXED.
6.3
At the time when I saw that the package was available for review, I was
writing an Archimedes version of Geordie Racer for Longman Logotron. I
was facing a serious dilemma, because the sounds I could easily produce
on the Archimedes were inferior to those which I had been able to create
for the original BBC model B version! SFXM sounded like just what I
needed, so I was on the telephone to Norwich Computer Services like a
shot!
6.3
The version which arrived a few days later is not the package I am
reviewing here, although the software itself is substantially the same.
I found the original documentation confusing and unhelpful. The copy
protection also caused some headaches before I realised that the review
copy had not been given the name that the protection routine was
searching for!
6.3
Nevertheless, it was clear from the examples provided, that this was a
highly versatile package capable of producing a tremendous range of
sounds and effects. The software was also very robust, and the ‘Sound FX
Editor’ (!SFXED) was the most versatile sound wave editor I had seen.
After some hours of frustration, I phoned C.I.S. It turned out that they
had received other adverse comments and were busy re-writing the manual.
6.3
After several months, the new version, with the documentation in draft
form, arrived. This time, the software protection routine (you have to
type in your name, and each disc is uniquely numbered) was completed
without fuss, and I was then able to transfer the software to my hard
disc.
6.3
The new manual is a vast improvement on the original so if you have the
old (32 page) manual, I suggest you get in touch with C.I.S. and request
an upgrade.
6.3
Making sound effects
6.3
So, how does it all work? Well, the basic idea is that you are given a
number of predefined and user-defined waveform and envelope generators,
plus some modifying tools like filters and delays, which you link
together to produce an output waveform. You can also start with any
sampled waveform and use the tools to modify it.
6.3
Everything is reasonably intuitive. There are a total of 16 ‘objects’
which generate or modify the waveform. These, and the wires which
connect them, are placed or dragged around the screen in typical RISC-OS
fashion. Alternatively, the different functions can be selected from the
usual RISC-OS menu structure but as this exactly duplicates the
facilities of the toolbox attached to the left of the sheet window,
there would seem to be little point. Saving a sheet is performed with
the usual save dialogue box which is selected from !SFXM’s iconbar menu.
6.3
When you have completed your ‘sheet’, you click on the calculate tool
and, after a delay which is never disturbingly long, the resultant wave
form appears in an !SFXED window and a ‘Keyboard’ window also appears:
6.3
6.3
At this point, you can hear what your waveform sounds like, as the
keyboard allows you to play notes in a three octave range. This can be
extended to eight octaves in conjunction with the function keys.
6.3
The waveform can now be further edited using the most comprehensive set
of editing tools I have seen. These include fade in or out; echo;
reverb; and silence, which is surprisingly useful for getting rid of
those annoying little blips which often seem to occur at the end of a
sample.
6.3
Samples can be saved in Tracker, Armadeus, DSEdit or EMR format.
Alternatively, the raw data can be saved or the sample converted to a
standard relocatable module. The sound sample can be converted between
linear signed and unsigned, and logarithmic form. The editor can, in
fact, be used alone − for example, to edit sound samples grabbed through
sampling hardware such as the Oak Recorder. It is not possible to undo
most of the effects but you can always get back to the original waveform
by recalculating it from the sheet.
6.3
Happily, this software comes with permission to use any samples created
in public domain or commercial software, without charge, CIS simply ask
for acknowledgement within the documentation or software which seems
very reasonable to me. (This contrasts notably with Oak Solution’s
approach to the software supplied with their Oak Recorder. I bought one
when they first came out, only to discover that the module supplied
which allowed the samples to be played was copyright!) (The solution is
to buy their Sound Lab software which produces relocatable modules − but
that costs a further £18 through Archive. Ed.)
6.3
My main complaint with the original manual was that it left me com
pletely in the dark about how I could use the package’s undoubted powers
to create my own sound effects in anything other than a completely
random trial and error fashion. The new documentation goes quite a long
way to addressing this problem, by giving examples of the way in which
the different generating and modifying tools can be used. The examples
are often cross-referenced to the 37 example sheets which are provided
on the disc. Also provided on the disc I received, were 25 sound effect
sample sheets, and 64 instrumental sample sheets. I understand that many
more than this will be supplied with the production version. C.I.S. are
currently working on a sampler board which will link to !SFXM and will
work through the serial port of the new A3010. (It will therefore work
on the A3020, A4000 and A5000, presumably. Ed.)
6.3
What I most want to be able to do, is to use the computer to create the
sound effect which I have in my mind. I do not believe that this is an
easy objective and I think I will go on experimenting for some time
before I am quickly able to get close to the effect I want. However, I
do think this is an excellent tool for learning about the generation of
sound, and with the new version of the documentation, this is by far the
best tool I have to help me.
6.3
The cost of this package is £49.95 from C.I.S. or £46 through
Archive. A
6.3
ClassPrint
6.3
David Leckie
6.3
Printing with a RISC-OS driver over Econet can cause severe problems due
to the limited bandwidth of this network. Any RISC-OS printer driver
produces a program for the printer in its own control language. Now,
with the exception of the PostScript driver, this will always describe a
bit image whether it be text or graphics that has to be printed.
6.3
This results in a very large amount of data having to be passed from the
client station to the network printserver. If more than about two
stations try to print simultaneously, the bandwidth of the network is
exceeded and “Not Listening” messages start to appear. If a class of
twenty want to print even a single A4 page from Impression, it can take
many hours, if not days, to achieve.
6.3
ClassPrint is an attempt to help overcome these problems by utilising
the speed of a Laser Direct printer at the printserver end and compress
ing the data at the client end. It has been produced jointly by Computer
Concepts and Oak Solutions though it is marketed only by the latter.
6.3
As Computer Concepts’ Laser Direct printers have previously been
reviewed in Archive 4.11 p2 this review will concentrate on the
facilities offered by the ClassPrint software. ClassPrint consists of a
small but quite adequate user guide and a single disk containing the
software. The software consists of Spooler, an application similar in
appearance to the Level 4 Spooler application. This is run on the
printserver along with PrinterHR and a modified versions of CC’s
PrinterHR is run on the client machine. Versions are supplied for the
LBP-4, LBP-8 and the now-discontinued Qume Laser Direct printers.
6.3
The hardware requirements are:
6.3
Client: Any A-series machine on Econet. Memory requirements depend on
the application that uses the printer driver. 1Mb is sufficient for
Paint/Draw, etc. Impression is tight on 1Mb so 2Mb is recommended.
6.3
Server: Any A-series computer connected on the Econet with 4Mb ram and
a hard disc with several Mb free. Though not recommended, this print
server can also function perfectly satisfactorily as a Level 4 file
server. An ARM3, while not essential, is certainly desirable. Laser
Direct printer which can be used simultaneously as a normal Laser Direct
printer and as a Laser Direct server.
6.3
For the following tests, the hardware/software configurations shown
below were used:
6.3
Client: 20 A3000s with 2Mb of ram and a 128Kb font cache configured.
Floppy drive only. Fonts etc taken from floppy not the network. New
Econet modules installed. RISC-OS 2.
6.3
Server: A310 with 4Mb ram, ARM3, Laser Direct hi-res card, Canon LBP-4
printer, 100 Mb Quantum SCSI drive with 50Mb free. RISC-OS2. The Print
Server was also running Level 4 file-serving software but no file-
serving demands were made during the timing tests.
6.3
Before installing the above system, the printserver software had been
Spooler (supplied with Level 4) and the printer an Epson LX400. The
older, but quite standard, 1.xx version of !PrinterDM was used by the
clients, at the highest resolution.
6.3
These sets of tests compare the time taken to print CC’s sample ‘News’
document from Impression II. The times shown below are not just until
control is regained by the client but are as recorded when the printout
is produced. Times are mins and seconds.
6.3
Printer To Spool Total to
Print File Size
6.3
Laser Direct N/A 0m 43s
272Kb
6.3
(as standard)
6.3
Laser Direct 1m 53s 2m 26s
271Kb
6.3
(as Print Server)
6.3
LX400 (1st client) 4m 53s 9m
17s 564Kb
6.3
LX400 (2nd client) 4m 53s
13m 45s 564Kb
6.3
With a dot matrix printer, printing can start before spooling finishes,
giving the 1st client a quicker printout. With subsequent clients the
queue has built up. The longer time is the more realistic.
6.3
It can be seen that using ClassPrint makes network printing a realistic
proposition. Using a dot matrix, at the very best, one printout every 14
mins is all that can be hoped for. In practice, not even this can be
achieved due to “not listening” messages when several stations try to
print simultaneously. However, there are still some problems with this
software.
6.3
Problems
6.3
• There is no way that a simple text printout can be produced because
!EmulateLQ, as supplied with Laser Direct, will not work over a network.
This means that software such as First Word Plus or any software that
uses a VDU 2 will not print. i.e. it must print via a RISC-OS driver.
PipeDream will work perfectly with the RISC-OS drivers but its own
drivers will not work. To get a listing/printout from a Basic program,
the output has to be spooled, tidied up in Edit then dragged to the
printer driver. Pascal source, prepared in Edit, is OK but sample run-
time needs to be spooled, tidied in Edit, then printed.
6.3
• The Spooler application from Level 4 will not co-exist with the
ClassPrint spooler. Thus, if some stations on the network need a simple
text printout, e.g. a listing, and others want the Laser Direct printer,
two separate print servers are required. Pupils become rapidly confused
when they try to select different print servers. (This problem may have
been overcome with AUN.)
6.3
• There is no method of putting a header with the client station’s I.D.
or station number on the printout. This can cause confusion in identify
ing pupil’s work, especially when all the work is similar.
6.3
• On occasions, an error message, “No room in shared area” appears. This
happens chiefly while printing with a long queue. Oak Solutions have
investigated this but believe that the problem lies with CC’s part of
the code. A re-write to overcome this problem is planned but is not
imminent.
6.3
• The client software is designed only to install if the printserver is
running. This is desirable but it can lead to the following problem.
When the client software is run, it searches the network, including
other networks across bridges, for a server called laser. It then checks
that the Laser Direct controller module is present in this machine.
However, included in the Level 4 Spooler application is a logical
printserver called Laser. (Presumably, this is to serve an HP Laser Jet
printer.) This could be anywhere on the network but, on occasions, it
may find this server before the correct one and will then, naturally,
fail to find the controller module and generate an error message. The
solution to this problem, once diagnosed, is easy − just rename the
Level 4 logical server to say Laser_3.
6.3
• Another similar problem occurs when two ClassPrint servers are run on
the same network. Again, while installing on the client machine, the
software searches the network for a server called Laser. However, there
is no way of knowing which server it will find. In our case, it
sometimes failed to find the server a few metres from it but found one
on another network two bridges distant. Printouts were thus directed to
the wrong server. This can be solved by editing the choices file inside
the spooler applications to say Laser1 and Laser2 as there is no menu
option to rename a print server.
6.3
• A large number of stations (10-20) still cannot print a large file
simultaneously without errors, but, due to the speed of the system, all
that is required is to spread the start of printing.
6.3
(Oak Solutions are aware of these problems and are working on the ones
that are caused by faults in the software. Ed.)
6.3
C onclusion
6.3
Laser Direct is a good product. It does what it is supposed to do and it
does it well. The real problem lies not with ClassPrint itself but with
the lack of bandwidth on Econet. It is a good attempt to overcome these
problems. In our case, it has transformed the horrendous problem of
network printing into a situation that, while not perfect, is adequate.
If the problems of simple text printing, the clients I.D. and the “No
Room in shared area” can be overcome, and if menu options can be
included to rename servers, ClassPrint will be an excellent product. The
guide should also include a “trouble shooting” section!
6.3
ClassPrint costs £79.95 +VAT from Oak Solutions or £87 through
Archive. A
6.3
ArcFS2 versus SparkFS
6.3
Tim Nicholson
6.3
Archive filing systems with some form of data compression have been
around for a while now, principally in the form of Compression from
Computer Concepts and ArcFS by Mark Smith. Enigma is due out soon which
promises data compression and encryption protection but in the meantime,
David Pilling has entered the scene with SparkFS, a filing system
implementation of his popular Spark archive utility, and an upgrade to
ArcFS in the form of ArcFS2 has been released by Software 42. Both these
packages are similarly priced and claim to offer similar facilities, so
how do they compare?
6.3
The packages
6.3
ArcFS2 comes on a single floppy disc together with some utilities,
including a PD read-only version of the application, and release notes.
It occupies 53Kb of disc space with 17Kb for the associated Info
application. The thirty-one page manual is either a good photocopy or a
bad DTP printout, and contains not only a tutorial and guide to basic
operation of the application, but also useful sections on such matters
as using the system with limited memory, differences between operation
with RISC-OS 2, RISC-OS 3 and RISC-OS 3.10, and a list of supported
*commands. The package is supplied in a video library case which is big
enough to let the floppy rattle around alarmingly and too small to take
the manual without folding it.
6.3
SparkFS comes on a single floppy together with the latest copies of
Spark and Sparkplug and the usual release notes, etc. SparkFS occupies
some 344Kb of disc space, of which 199Kb is recoverable on your working
copy if you are content to limit the range of archive formats which you
wish to handle. It is packaged in a purpose-made vinyl-backed wallet
together with a “properly” printed manual of 28 pages which includes a
history of the various archive formats SparkFS understands, a descrip
tion of supported *commands and a list of error codes/messages.
6.3
The facilities
6.3
ArcFS2 provides an archive filing system utilising file archives of
filetype_DDC. It supports various compression algorithms, none, Pack
(where only consecutively repeated bytes are compressed) or 12 to 16 bit
LZW compression of the Crunch or Compress variety. Each has its own
merits of speed and memory requirements. “Garble” password protection is
available and this is independently settable for different objects
within one archive. Improvements over the original version include speed
and the number of archives “open” at any one time. A separate utility
called ArcFSInfo is provided to enable archive statistics to be viewed.
This utility has its own icon on the right of the iconbar which is in
addition to the ArcFS2 icon on the left(filer) side of the iconbar.
6.3
SparkFS likewise provides an archive filing system but may utilise a
variety of archive formats including (currently) those found on other
hardware platforms including PC’s and the Mac. I say currently because
SparkFS is written as a series of modules, each one designed to handle a
particular format, and a core filing system module which is format
independent. Thus if a new format comes along, upgrading is simply a
question of obtaining a new format specific module. Also, formats which
are currently read-only may be upgraded to read/write if there is
sufficient demand to make it worth the author’s while. Currently
supported archives are, read-only:- ARJ, LZH, McStuffitt, PackdDir and
Zoo. Read and write:- Spark files_DDC, Spark directories (a new animal,
similar to CFS directories), PK arc, Zip and Tar. For those who have
never heard of half of these, the manual gives a short potted history.
For those who never need to know, the irrelevant modules can be deleted
from the working disc, saving up to 199Kb of space.
6.3
SparkFS also supports a large range of file conversion facilities,
primarily of use to those wishing to send 8bit data down 7bit channels
(e.g. viewdata systems). Of the various standards around, SparkFS can
handle uucode (from Unix), atob, FCET, boo and HQX. Again, the manual
explains the various formats to those unfamiliar with them. Compression
options depend upon archive type but for Spark file or directory
archives, they are:- None, Squeeze, Squash, 12 bit Crunch and 12 or 16
bit compress. Encryption may either be “Garble” or DES, which is
supposedly a very secure algorithm, as used by NASA, and may be applied
to individual objects within an archive. Archive statistics are obtained
by control-dragging an archive to the filer icon rather than needing a
separate utility, but are not as comprehensive as ArcFS2’s.
6.3
In use
6.3
A double click loads ArcFS2 on the left (filer side) of the iconbar.
Thereafter any archives can be treated just like any other part of the
filing system. A preferences window can be selected on the iconbar for
setting encryption, compression type and miscellaneous options. These
preferences may be saved as the default settings or used “just for now”.
The iconbar menu allows creation of new archives, manual compaction of
existing archives which have had files deleted from them, and facilities
for opening the root directories and parent directories of open
archives, or closing archives and, finally, quitting the application.
6.3
Once an archive has been opened, ArcFS2 keeps a note of it. If the
window is closed it may be easily reopened or the directory in which the
archive exists reopened using the Open $/parent facilities in this menu.
This is a useful way of avoiding cluttering up the desktop with too many
open windows. ArcFS2 refers to archives by their leafname and thus will
not allow two archives of the same name but different paths open at the
same time. Also, if an archive is deleted or moved but not closed first,
it will still appear in the list of open archives. Attempting to Open $
produces a window showing the “contents” of the archive, but attempting
to use the contents gives a ‘file not found’ error. This is slightly
annoying and can lead one to think one has archives which no longer
exist! When an archive has been created it is not automatically opened.
This must be done by double clicking on it.
6.3
Since ArcFS2 is written as a module, it uses the module area for its
workspace. The manual warns of the need to ensure enough free space in
this area and gives a table of memory requirements and the procedure to
adopt depending on whether RISC-OS 2 or RISC-OS 3 is being used.
Contrary to the instructions in the manual, I had to adjust the module
area memory allocation to avoid running out of memory with large files
even with RISC-OS 3. This became an annoyance when it happened halfway
through an operation and I had to start again.
6.3
A separate utility, ArcFS2Info, is provided to show statistics of an
archive dragged to it. It gives information about original file size,
compressed file size, compression type, read/write status and compres
sion ratio (compressed file size as a percentage of original). Whilst
this is a useful utility, I think it is a shame that one needs another
icon on the iconbar (mine is always too full anyway).
6.3
SparkFS’s iconbar menu is shorter than ArcFS2’s containing only Info,
New Archive, Choices, and Quit. Choices is the equivalent of ArcFS2’s
select option on the archive bar and leads to a preferences window
containing options for archive type, compression type, default temp
file, archive modules loaded, max memory SparkFS may grab and options
relating to UUcoded files. Encryption settings are dealt with in the
statistics window opened by Control-dragging an archive to the iconbar.
6.3
One big difference between the two applications is the use of workspace
memory. As noted, ArcFS2 uses the RMA, while SparkFS is configurable.
One option uses the system sprite area, others use memory claimed from
the operating system. SparkFS will use as much memory as it needs, up to
its allowance as set in ‘max’ in the choices window. If that is not
enough, it will use a Temp directory which, as supplied, is set to
<Wimp$ScrapDir> but this may be altered to any convenient location
including RAMFS. Using a Temp directory does slow things down a bit but
it means that SparkFS can handle big files more readily than ArcFS2
without running out of memory.
6.3
The New archive option allows one to specify the type of archive and
name and then it is created in the usual way of dragging to a filer
window. New archives are automatically opened after creation. Clicking
<select> on the iconbar provides a short cut to this menu option.
6.3
Spark file archives are compatible with those of Spark and similar to,
but not entirely compatible with, ArcFS2 files. Spark directories are a
new form of archive similar in structure to CFS directories. They are a
pseudo-application with a Data directory containing the actual archives
as individual objects. The default Sprites file is the familiar
lightning flash on blue but can be customised to whatever you fancy to
aid the identification of an archive.
6.3
Directories are more robust than files since each item is a discrete
entity. If files are deleted, the archive does not need compacting and
recovery should be easier after a disc fault. The down side is a
slightly bigger overhead reducing the compression efficiency. Once an
archive has been created, there is no necessity for the iconbar front
end of SparkFS and the Quit option allows for quitting of the front end
only or of the Filer too.
6.3
The lack of Open $/parent is not really a problem with RISC-OS 3 since
the Pinboard facility can be used to keep the number of open windows
under control whilst keeping archives accessible. As with ArcFS2,
archives may be opened by double clicking on them or dragging them to
the iconbar. Shift-dragging performs conversion operations, the options
depending upon the source object type. Spark files can be converted to
Spark directories and vice versa. Applications and directories can be
archived directly without the need to create the archive first and
binary to text (and back) conversions are handled in this way.
6.3
Performance
6.3
In order to compare the speed of the two packages, I set them to work on
a Squirrel database of 1,352,716 bytes which I know to be readily
compressible. The reason for the choice was twofold. Firstly, it enabled
me to do timings with a stopwatch without introducing significant
reaction timing errors and secondly it would show up any weaknesses in
the memory management of the applications. It is a real use which I do
weekly, at least, and sometimes daily. For the test, I used a standard
ARM2 A440 RISC-OS 3.10 with ST506 hard disc, creating the archive/
unpacked data on a RAM disc from the source object on the hard drive.
This was designed to keep hard disc access down to a minimum and so
reveal the speed of the processing algorithms. The compress 12 algorithm
was used in all cases and the compressed archive size was about 30% of
the original data. (The time taken to copy the file directly from the
hard drive into the RAMFS was 9.5s.)
6.3
Spark SparkFS ArcFS2
6.3
create archive ~120s 38s (1)
40s *
6.3
46s (2)
6.3
56s (3)
6.3
unpack archive ~90s 22s (1)
38s *
6.3
33s (3)
6.3
Notes: (1) using memory mode, max dragged out to
accommodate largest file
6.3
(2) using Temp Directory in RAMFS:
6.3
(3) using <WimpScrap$Dir> on :4
6.3
* initially ArcFS kept running out of memory in RMA until I used the
task manager and dragged it out to as much as I could. This is not as
the manual says it should be under RISC-OS 3.
6.3
All filer operations were verbose but in the case of option (1) did not
actually appear for reasons I am unsure of. It is interesting that
almost as much time is saved switching from RAMFS to memory as from
<WimpScrap$Dir> to RAMFS which must indicate the filing system overheads
in read write operations even if they are to RAMFS. The saving is of the
order of the raw transfer time from :4 to RAMFS! Subtracting the raw
data transfer times from the above figures gives a processing time of
around 30s to pack and 20s to unpack (30s for ArcFS2) around 1.3 Mb of
data. An ARM3 would presumably reduce these times further. This means
that for more modest operations, the processing overheads will not add
unreasonably to disc operations, e.g. unpacking a 100Kb file should only
take an extra couple of seconds.
6.3
Compatibility
6.3
ArcFS2 will happily read Spark File archives created by Spark or SparkFS
but it declares them as read-only and will not add to them. A utility to
covert them is supplied with ArcFS2 or they may be simply dragged to a
new archive. It cannot read Spark directories.
6.3
SparkFS declares ArcFS2 archives as read-only if one attempts to write
to them. Attempting to read the archives caused SparkFS to crash with
memory errors and trying to quit SparkFS sometimes caused the computer
to hang up. This seems to be because the Spark module has got corrupted
and will not die when asked. This is a problem because, when faced with
a file_DDC archive, there appears to be no way of knowing which
application created it. For ArcFS2 this is not a great problem but for
SparkFS it could cause an embarrassing crash. Apart from this problem
SparkFS can handle a great many more formats than ArcFS2 so it would be
nice if this little hiccup could be fixed. It is interesting to note
that the two archives created by the two applications from the same
source were different sizes, the ArcFS2 file being some 1580 bytes
bigger.
6.3
Conclusion
6.3
These two packages are similarly priced. For my usage, the memory
management options of SparkFS compared with ArcFS2’s limitations on
handling large files make it my first choice. For more modest use, this
would not be a problem. However, there would still be the other areas
where SparkFS scores over ArcFS2 − no separate icon cluttering up my
iconbar just for archive statistics, not limited to one archive format
only, and all the file conversion options. I would like to have a read-
only PD version that I can leave on the machine at work.
6.3
ArcFS2 is smaller and, for those with problems of space, this may be
important but this difference is not great if one cuts down on the
archive options of SparkFS. ArcFS2 comes with a read-only version which
is PD, there is no equivalent in SparkFS and Sparkplug is getting long
in the tooth now. It is incredibly slow by comparison and can’t handle
directory archives. Apart from the distribution aspect of providing a
read-only version, there is an element of protecting software against
unauthorised tampering, obtained by using archived applications and a
read-only de-archiver, so how about it, David Pilling?
6.3
From the initial presentation of the packages, through to the facilities
provided and the way they are implemented, SparkFS comes out a clear
winner for me. ArcFS2 is a notable improvement to the earlier version
and those familiar with it may prefer to stay with the style of
something they are used to, whilst gaining the benefit of the enhance
ments. Both packages come with a discount for those upgrading, but
SparkFS comes with a double discount for those with both ArcFS and
Spark. This can make it even better value for money. For those with
twice as much to spend, there is still Computer Concepts CFS and, round
the corner, is Enigma (with what facilities? and at what cost?) but, at
the moment, SparkFS wins on price and performance.
6.3
ArcFS2 is £20 (no VAT) from Software42 and SparkFS is £23.96 from David
Pilling or £23 through Archive. A
6.3
Rheingold Enterprises 17 Ingfield
Terrace, Slaithwaite, Huddersfield HD7 5BJ. (0925−210657)
6.3
Risc Developments (pp6/10) 117 Hatfield
Road, St Albans, Herts AL1 4JS. (0727−40303) (0727−860263)
6.3
Sherston Software Swan Barton,
Sherston, Malmesbury, Wilts. SN16 0LH. (0666−840433) (0666−840048)
6.3
Software 42 109 Ferry Road, Hullbridge, Essex SS5 6EL.
6.3
Soft Rock Software 124 Marissal
Road, Henbury, Bristol BS10 7NP. (0272−761685)
6.3
Spacetech (p32) 21 West Wools, Portland, Dorset DT5 2EA.
(0305−822753) (0305−860483)
6.3
State Machine 75 Upper Wellington Street, Luton, Bedfordshire LU1
5AA. (0582−483377) (0582−480833)
6.3
Techsoft UK Ltd (p9) Old School
Lane, Erryrs, Mold, Clwyd CH7 4DA. (082−43318)
6.3
Turing Tools 149 Campbell Road, Cowley, Oxford OX4 3NX. (0865−775059)
6.3
6.3