home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Atari Mega Archive 1
/
Atari Mega Archive - Volume 1.iso
/
lists
/
gem
/
l_1199
/
1084
< prev
next >
Wrap
Text File
|
1994-08-27
|
6KB
|
167 lines
Subject: app_defs
Date: Thu, 28 Jul 1994 12:01:59 +1000
From: Warwick Allison <warwick@cs.uq.oz.au>
Precedence: bulk
Ofir:
>[re: other attributes than shortcuts in the app_defs file]
>
>I think the best thing is to do a stage at a time. It's fine setting
>zillions of standards, it's another thing getting them implemented. The
>more we do the less support we get.
>
>The first app-defs file should only deal with keyboard shortcuts. The
>file format should be open so that we could add more things at a later
>stage.
Choices we make for the names of attributes in the app_defs file don't
require any implementation: they're just choices so as to have commonality
for any apps that want the same attributes. It's like choosing a cookie:
doesn't really matter what it is, provided everyone knows.
Summary of the issues I can think of:
1. Format (eg. see below)
2. Significance of order of entries (eg. does later override former?)
3. Location (eg. C:\SYSTEM)
4. Attribute names (eg. keyOpen? shortcutOpen? etc.)
5. Value representations (eg. ^C? Ctrl-C? Shift-X? 0x6643?)
6. Attributes other than shortcuts.
Inlike the short-cuts file, I think we should be able to come to agreement
pretty easily on this, as (1) it is fairly simple to implement (2) there is
no real existing standard [do we count assign.sys :)] (3) it doesn't involve
many concerns for the user [as Ofir said - many of them will never touch it].
Re the format,
It may be useful to have a `type' code for each atrribute-value pair in
the app_defs.sys file, to facilitate editing via a tool. For example:
*.papyrus*.key.openFile: KEY = ^O
*.papyrus*.font: FNT = Gemini 10pt
*.papyrus*.key.newFile: KEY = 0x00380402=F1 // Pretend this is the F1 code
*.papyrus*.fgColor*: COL = Purple
*.papyrus*.bgColor*: COL = 0x00150315 // Pretend this is a nice colour
*.tempFile*: FIL = f:\tmp
*.foo.name*: TXT = Hello World
<attribute-pattern> ":" <type> "=" <value>
Anyone not like the ":" and "="? They're redundant anyway:
<attribute-pattern-no-spaces> <whitespace> <type-no-space> <value>
The type need not be 3 characters.
A app_defs.sys editing tool could then know how what to do with ALL
lines as far as allowing the user to edit them. It also helps the user
to know what they can put for an option. Applications reading the
app_defs.sys file could either ignore the type, or just check that it
is as expected. Only one type should be allowed for each attribute,
but different forms of representing the value of that type in text may
be allowed (such as the text "^O" and the scancode+menuname "0x00380402=F1").
All apps should support the same conversions from value text to actual
value, so the choices should be kept pretty small.
The format of an attribute could be:
<application-type>.<application-name>.{<window>|<other qualifier>}.<name>
It doesn't really matter what the format is, as all an application needs
to do is a wildcard match against it.
Standard application-types could be:
word
image
dtp
util
game
The purpose of the application-type is to allow class-wide attributes,
such as:
game.*.hog_cpu: true
It's just an option - I'm not dead-set on the idea, but it costs us nothing,
since we can just pattern-match anyway.
David Oakley:
>For keys, I think we need to formalise as many keys as are used in a wide
>number of applications. This means all of the keys that Atari/GemList
>specify, plus a few more if we can think of them.
That's a good idea. If only to define the names of the attributes. eg, I
might have:
*.nethack*.key.saveAndQuit: Shift-S
But I don't expect us to make a `standard default' for saveAndQuit.
There are many others. Basically, we can just take a look at all the
menu items in all the programs we have and divide them into 4 groups:
1. standard (eg. Open)
2. not standard (eg. Load)
3. common (eg. Goto Mark)
4. not common (eg. Offer Corpse To God)
(1) and (3) should have names chosen for them.
(1) should have standard defaults decided for them (which we've almost done).
(2) should be listed in gemlist docs as Do Not Use, along with proper term.
(4) can be ignored.
>Dialogs in Windows (on/off)
>Novice User Help (on/off)
>Confirm on Exit/Save on Exit
>Default Window Size
Here's some *possibilities* (not sure they're that useful):
WindowGadgets (MOVER|...)
WindowGeometry 80x25+10+10
SizingProtocol (click-move-click / drag)
MovingProtocol (click-move-click / drag)
PopupProtocol (pop-up / drop-down / ...)
User.Handedness (left / right) // Hey! It' might be useful :)
Keyboard Focus (mouse/top/click) Ouch, Ofir, stop hitting me! Ouch! Oooof.
(just to give it a standard name - I'm not saying people should use it,
just that if they do, this is how it is enabled)
>Auto Save Prefs on Exit
This is an interesting one. Will there BE any `preferences' dialog, once
we have an app_defs.sys file and a GUI for editing it? Some of the issues
are: do you want to set window positions outside the application? should
an application modify the app_defs.sys file on a `save preferences' operation?
Note that the above can still all be configured on a per-application basis,
so even if you generally hate `Confirm on Exit', there might be a couple of
applications for which you enable it.
BTW, I volunteer to make an app_defs.sys editor. (of course, it will be 500K ;)
chrisridd:
>What we *should* be doing, IMHO, is proposing and voting for the
>existence of a shortcut file or files, and only then discussing the
>precise contents of it. The whole voting period should be much shorter
>too, as this list is very active :-) Say two weeks per vote.
Voting before discussion? That's not very democratic. It's like voting
for a parliament, THEN letting them discuss policy.
I've certainly changed my mind about a lot of things since listenning
to the other list subscribers. Preclude discussion and you end up with
the type of standards comittees you often get in industry: where
everyone is just trying to make THEIR PRODUCT the standard, regardless
of its merits.
--
Warwick