Simtel MSDOS 1992 December/simtel1292_SIMTEL_1292_Walnut_Creek.iso/msdos/database/tara201.arc/TARA.DOC

Text File | 1987-05-12 | 25KB | 569 lines

              Tara Datafile Utilities, Version 2.0

                          USER'S MANUAL
                               by
                         David C. Oshel




              Copyright (c) 1987 by David C. Oshel

                       ALL RIGHTS RESERVED

Private  individuals  are  granted  free  license  to  copy   and
distribute this complete set of Tara Datafile Utilities, provided
my  copyright notice is not removed, and provided you  distribute
both programs and documentation without charge.

Corporate  or governmental use requires a license fee of  $25.00.
Send check or money order payable to:

                    MicroConsulting Services
                       1219 Harding Avenue
                         Ames, IA 50010



I.  Introduction

These files are included in Tara Datafile Utilities, version 2.0:

File             Purpose
-------------------------------------------------------------------------
BROWSE.BAT       demonstration, uses STARS.DAT
CRYPTIC.DOC      (doc)
CRYPTIC.EXE      sophisticated file encipher, decipher
ENTER.DOC        (doc)
ENTER.EXE        data entry for MailMerge-type datafiles
FIELD.DOC        (doc)
FIELD.EXE        selected fields, with format for browse
FNKEY.DOC        (doc)
FNKEY.EXE        assign macro strings to PC function keys
NAMES.FLD        sample field definition file for NAMES.DAT
NSORT.DOC        (doc)
NSORT.EXE        numeric sort on any field or fields
PICK.DOC         (doc)
PICK.EXE         select records on any field or fields
SETNAMES.BAT     define Fn keys for use with NAMES.DAT
SETSTARS.BAT     define Fn keys for use with STARS.DAT
STARS.DAT        demonstration data file, list of bright stars
STARS.FLD        field definition file for STARS.DAT
TARA.DOC         you're reading it
TSORT.DOC        (doc)
TSORT.EXE        text sort on any field or fields


The  Tara Datafile Utilities are a set of small,  powerful  tools
which were originally intended just to make life with MailMerge a
little  easier.  Until now, information retrieval from this  kind
of data file meant you had to use WordStar's Merge Print facility
to restrict, select out, and display your data.  If you wanted to
sort (or resort) your data before using it, you were out of luck.

But with these tools, you now have direct access to your own data
and you no longer need to fire up WordStar and a printer just  to
see what's what.

Most of these programs are simple, one-task tools.  They are used
in  combination with each other and with the MS-DOS command  line
i/o  redirection facility.  In general, Tara  Datafile  Utilities
allow you to:

a)  restrict  the mass of data you  have to the few  records  and
fields you actually need to look at, based on the contents of any
field or combination of fields;

b)  sort  this smaller subset of information in either  ascending
or descending order, on either text or numeric data, on any field
or combination of fields;

c)  project  the results to another file for use by  WordStar  or
some other program, or to the screen, where you may view selected
and/or formatted fields at leisure.

Tara Datafile Utilities also include:

d) a quick data entry module, using a field definition file which
you create;

e) a sophisticated file encryption program;

f) a program which assigns macro strings to PC function keys.


                               ###


II.  MS-DOS i/o redirection and pipes

You  should  have a firm grasp of what "i/o redirection"  is  all
about  in  order to use the Tara Datafile Utilities.   (There  is
also a brief discussion of this issue in your MS-DOS manual.)

In  general, "i/o" means INPUT TO and OUTPUT FROM any  particular
program.  Most commonly, input comes from your computer  keyboard
and output goes to your computer screen.

Less commonly, input may come from a text file, and the stream of
characters  from that file is treated exactly as  though  someone
were rapidly typing on the computer's keyboard.

Similarly, output may be sent to some other destination than  the
computer screen -- either to a line printer or to a text file.

Less commonly still, the OUTPUT from one program can be the INPUT
to another program!  This is called a "pipe".

So,  the three things to be aware of are 1) redirected input,  2)
redirected output, and 3) pipes.  These three demons are  invoked
on the MS-DOS command line, and nowhere else.

The  DEFAULT input is "<CON:" and the DEFAULT output is  ">CON:".
You  do  not need to specify the DEFAULT input or output  on  any
command line.  However...

Input from a FILE has the form "<file1.dat" and output to a  file
has  the  form ">file2.dat".  Output to the  line  printer  would
typically be written as ">PRN:", but you may also see ">LPT1:" or
">LPT2:" in the case of serial printers or modems.

A  PIPE  is  specified  with the  vertical  bar  character,  "|",
surrounded by spaces fore and aft, " | ", between two COMMANDS:

                       C>dir | sort | more

This example comes entirely from MS-DOS; i.e., "dir", "sort"  and
"more"  are all MS-DOS commands, and they are discussed  in  your
MS-DOS manual.

Note that MS-DOS sort typically REQUIRES i/o redirection:

                  C>sort <inputfile >outputfile


Note also that the SOURCE file and the DESTINATION file must  NOT
be the SAME file!  I.e., this is a serious error which  typically
destroys your source document:

           *** DANGER ***  C>sort <abc.dat >abc.dat


                               ###

III.  Features common to all Tara Datafile Utilities

A.  Record type is MailMerge-compatible

All  Tara Datafile Utilities assume that the data they work  with
is  compatible  with  MicroPro, Inc.'s  WordStar  program  --  in
particular with Merge Print (a.k.a. MailMerge in older versions).

That means, in general, that records consist of a single line  of
characters  terminated  with  a carriage  return/line  feed  pair
("newline"), and that fields within each record are delimited  by
commas, except for the last field.  If a field contains a  comma,
the  field  is enclosed in double quotation marks.   If  a  field
contains  both  a  comma and quotes, the  field  is  enclosed  in
apostrophes.   If neither of these quotation schemes is  adequate
to  mark off the fields in the record, the  delimiting  character
can be changed from comma to something else.

For example:

    "Oshel, Ph.D.",David,C.,1219 Harding,Ames,IA,50010,,,yes

There  are  THREE  fields following the zip code  comma  in  this
example.   The last field is terminated by the same newline  pair
that terminates the record.  The first field contains a comma, so
that field is quoted.

Tara Datafile Utilities refer to the fields in a record by  FIELD
NUMBER.  The nth field in each record is to the immediate left of
the nth comma (or chosen delimiting character).  In the  example,
the  word  "Harding" lies in the fourth field, there is  no  home
phone number in the eighth field, and the word "yes" occupies the
last, tenth field.  The ENTER program provides field numbers  for
ready  reference -- you don't HAVE to count your commas!  You  do
have to refer to fields by the numbers.

There is NO DISTINCTION between upper or lower case in any of the
Tara Datafile Utilities.  This applies generally and  everywhere.
However, upper and lower case IN DATA are preserved when data  is
written to a new file or to the screen.

All of these utilities STRIP THE HIGH BIT out of characters  when
writing data to standard output.  This does not alter the  source
data, but does produce a file which contains no "negative  ASCII"
characters -- on IBM PC's, these are the letters with diacritical
marks, box characters, greek letters, etc.  WordStar uses them to
hide formatting information in DOCUMENT MODE text.

This  fixup allows PICK, TSORT and NSORT to work with  data  that
was  inadvertently  edited in WordStar document  mode.   You  can
detect  "funny  letters" in your data by using  the  MS-DOS  TYPE
command  to examine your file.  Negative ASCII  characters  would
prevent  the  pattern-matching utilities from  finding  "obvious"
matches.  Tara Datafile Utilities do not alter your source  data,
unless you write it back to the original file through a pipe.

Notice that MailMerge-type datafiles are somewhat  idiosyncratic.
Another useful definition of an "ASCII data file" is that  fields
with  text data are quoted, fields containing quotes  double  the
quotation character (e.g., """" defines a field which contains  a
single  "  character!),  but numeric data is  not  quoted.   Some
programs  that support some version of this other definition  are
BASIC and R:Base 5000.  The difference lies in how to handle  the
ambiguity caused by quoting the quote character.  WordStar  tries
to duck the issue by adding ' as another quote character.

However, there is no hard and fast rule or widely held  standard.
These  other formats are chiefly used only to import  and  export
data  to and from expensive data base management programs,  which
do not themselves use the format internally.  "MailMerge  format"
is extremely and immediately useful (to WordStar as well as other
programs) so that is the format chosen here.  Data imported  from
R:Base  5000  or dBase III ("DELIMITED BY ,")  will  probably  be
acceptable to Tara Datafile Utilities, and probably acceptable to
WordStar  Merge  Print, but you may need to do  some  preliminary
fixing  up.   (If  you,  like me, have  either  of  those  pricey
database programs you're probably using Tara for the same  reason
I  do  -- Tara is quick and dirty and easy to live with.   But  I
wouldn't do a fancy job with just these simple tools.  Yet...!)

This kind of record has one distinct advantage -- it is  variable
length -- and one obvious disadvantage -- it takes longer to  get
information out of the file (especially numeric information).  In
general, MailMerge-type data files should not contain more than a
a  few hundred records, or processing time will be tedious.   But
you can always PICK a smaller set of data from a larger file.

The  largest record Tara Datafile Utilities can  handle  contains
4095 characters (with any number of fields within that limit).

                               ###

B.  Switches

The "-H" or "/H" switch always invokes a help screen for each  of
the utilities.

Examples:     C>enter -h
              C>pick /h
              C>fnkey -h

The  "-F"  or "/F" switch changes the field  delimiter  character
from comma to something else, in those utilities that read data.

Examples:     C>pick -f\ <abc.dat 1 has aardvark
              C>field <temp /f* /s. 1 12 3 8: | more
              C>nsort "-f|" 16 d

But:          C>fnkey -f9 "pick -f\ <abc.dat"

The "-S" or "/S" switch only occurs in FIELD.EXE, and is used  to
change  the  fill  character from BLANK to  something  else  when
right- or left-justifying formatted output.

The "-E" or "/E", and "-D" or "/D", switches are used in  CRYPTIC
and nowhere else, to indicate mode:  encipher or decipher.

All program switches begin with either hyphen or slash.  Switches
alter  the usual behavior of the utility in some way.  There  are
only one or two switches at most for each utility, in addition to
the help switch.

Undefined switches usually, but not always, invoke the  utility's
help  message.  FNKEY will not try to interpret any switches  but
its own, viz., -H and the first instance of -F.  Negative  number
arguments are never interpreted as switches.

Note that a switch which contains <, |, or > must be QUOTED.


                               ###

C. Standard Output

All of the utilities support redirected i/o using standard  input
and standard output.  This feature is not especially useful  with
ENTER or FNKEY but it's there if you can figure out a use for it.

("Standard input" and "standard output" refer to the method which
a  program uses to receive or transmit data.  The "standard"  way
to  do  things supports MS-DOS i/o redirection  and  pipes.   The
various  non-standard i/o schemes used by almost  all  commercial
programs are much faster ... and also much less flexible.)

                               ###

D.  ANSI.SYS, only for FNKEY.EXE

None  of these utilities need to have ANSI.SYS installed on  your
system  -- except for FNKEY, a program that assigns  meanings  to
your PC's function keys.  See your MS-DOS manual for instructions
on installing ANSI.SYS.  What you need to do is place the command
DEVICE=ANSI.SYS into the CONFIG.SYS file on your startup disk.


                               ###

E.  Numeric and dollar data types

Utilities  that recognize numeric data will  correctly  recognize
the  dollar format, e.g., $123,456.78.  A minus sign  may  appear
anywhere in its proper scan field.  All numeric data is converted
to floating point for purposes of comparison.  Scientific formats
like "1.3e-2" are supported, but "2e" is not considered a  number
(because  there is no digit following the E); the  exponentiation
operator  may be either E or D.  Scientific and dollar  notations
are  mutually incompatible in the same field.  The FIELD  utility
relaxes the strict interpretation of what is a number, and allows
multiple hyphens, parentheses and slash -- this allows data  like
social  security numbers and phone numbers to be  right-justified
in formatted output.  All non-numeric data is Text.


                               ###


IV.  The CRYPTIC Program

Usage:  C>cryptic { -D | -E } password <inputfile >outputfile

Examples:   C>cryptic absinthe <myfile.dat >coded.dat
            C>cryptic -D absinthe <coded.dat | more

The  -D  switch  selects Decipher Mode.  The  -E  switch  selects
Encipher Mode, and is the default mode.  You must supply the same
password each time you use Cryptic with a particular file.

         *** WARNING:  Do NOT forget your password! ***

This  is a simple "filter" which scrambles the contents  of  text
files.   Coded  files  are indecipherable by  normal  means.   To
decode  the  scrambled file, run it through  CRYPTIC  once  again
using the same password as before.

If  you  forget  your  own password, you're  out  of  luck.   The
password  is  not recoverable.  Remember it!  You should  keep  a
backup copy of any data file you use, in your possession and  off
the  premises.  This anticipates the problem that  someone  might
maliciously encrypt your data for you.

This  utility provides a first level of data security  only.   It
will prevent unauthorized access by average persons, but will not
withstand expert analysis.

Be  advised  that your physical disk medium probably  retains  an
image  of some part of your plain text, even if you  have  erased
the file, unless you reformat the disk (after copying your  coded
data to another disk, of course!).

Using  a PIPE with plain text in it will also leave  a  transient
image  on physical disk media!  If security is crucial, use  your
floppy  drive and be sure to format the working diskette after  a
session.  This physical data image, not associated with any file,
can be viewed (and security compromised!) by any number of garden
variety programs including Norton Utilities and PC-Tools.

The encryption algorithm used here is more sophisticated than the
usual "xor" type of scramble.  In its day, this cipher could  not
be  cracked,  but no doubt things have changed a  bit  since  the
Crimean War.

To  illustrate the potential difficulty of cracking this  cipher,
Tables 1 and 2 compare the frequency of occurrence of  characters
found  in a plain-text data file, against the frequency found  in
an especially cryptic version of the same data.

The enciphered data file was run through Cryptic four times using
a different password each time.  Ciphered data has "smeared" over
the range of all printable characters, while the gap between most
and least frequent characters is far less than in the plain text.

Both the plain text and the ciphered data contain 117 records and
a total of 8,602 printing characters.  Tables follow.


Table  1.   Frequencies of 79 characters found in  a  plain  text
            data file delimited by commas.

        , = 1088, 12.6482%             R = 48,   0.5580%
          = 563,   6.5450%             k = 46,   0.5348%
        e = 532,   6.1846%             v = 46,   0.5348%
        r = 339,   3.9409%             P = 41,   0.4766%
        o = 335,   3.8944%             H = 40,   0.4650%
        a = 314,   3.6503%             f = 40,   0.4650%
        0 = 302,   3.5108%             O = 40,   0.4650%
        n = 275,   3.1969%             ) = 37,   0.4301%
        s = 268,   3.1156%             ( = 37,   0.4301%
        t = 260,   3.0226%             b = 34,   0.3953%
        2 = 251,   2.9179%             * = 32,   0.3720%
        i = 241,   2.8017%             L = 30,   0.3488%
        1 = 207,   2.4064%             U = 28,   0.3255%
        A = 204,   2.3715%             W = 27,   0.3139%
        l = 173,   2.0112%             J = 24,   0.2790%
        5 = 170,   1.9763%             G = 23,   0.2674%
        3 = 168,   1.9530%             N = 23,   0.2674%
        m = 139,   1.6159%             E = 22,   0.2558%
        C = 138,   1.6043%             & = 20,   0.2325%
        d = 129,   1.4997%             T = 19,   0.2209%
        4 = 122,   1.4183%             x = 17,   0.1976%
        - = 119,   1.3834%             F = 16,   0.1860%
        I = 118,   1.3718%             ? = 16,   0.1860%
        S = 110,   1.2788%             / = 14,   0.1628%
        9 = 104,   1.2090%             K = 12,   0.1395%
        8 = 104,   1.2090%             V = 11,   0.1279%
        c = 101,   1.1741%             Y = 10,   0.1163%
        u = 100,   1.1625%             # = 7,    0.0814%
        h = 94,    1.0928%             z = 6,    0.0698%
        6 = 92,    1.0695%             ; = 4,    0.0465%
        7 = 88,    1.0230%             : = 2,    0.0233%
        p = 79,    0.9184%             ' = 2,    0.0233%
        M = 69,    0.8021%             X = 1,    0.0116%
        . = 66,    0.7673%             ! = 1,    0.0116%
        y = 66,    0.7673%             Z = 1,    0.0116%
        D = 62,    0.7208%             j = 1,    0.0116%
        " = 62,    0.7208%             q = 1,    0.0116%
        B = 58,    0.6743%             \ = 1,    0.0116%
        g = 58,    0.6743%             Q = 1,    0.0116%
        w = 53,    0.6161%

Table 2.  Frequencies of 95 characters found in a 4-ply cipher of
          the same data.

        0 = 155,   1.8019%             p = 90,   1.0463%
        . = 131,   1.5229%             K = 89,   1.0346%
        6 = 129,   1.4997%             m = 88,   1.0230%
        , = 127,   1.4764%             w = 88,   1.0230%
        % = 123,   1.4299%             D = 87,   1.0114%
        ' = 123,   1.4299%             H = 87,   1.0114%
        # = 122,   1.4183%             e = 86,   0.9998%
        2 = 121,   1.4066%             g = 86,   0.9998%
        * = 121,   1.4066%             5 = 85,   0.9881%
        + = 119,   1.3834%             o = 84,   0.9765%
        v = 118,   1.3718%             M = 83,   0.9649%
        x = 117,   1.3601%             ^ = 83,   0.9649%
        ( = 116,   1.3485%             G = 83,   0.9649%
        | = 116,   1.3485%             n = 81,   0.9416%
        8 = 115,   1.3369%             > = 81,   0.9416%
        3 = 114,   1.3253%             ; = 80,   0.9300%
        4 = 113,   1.3136%             { = 79,   0.9184%
        ! = 113,   1.3136%             J = 79,   0.9184%
        ~ = 112,   1.3020%             Q = 78,   0.9068%
        } = 111,   1.2904%             N = 76,   0.8835%
        " = 110,   1.2788%             X = 76,   0.8835%
        = = 109,   1.2671%             a = 75,   0.8719%
        r = 109,   1.2671%             f = 75,   0.8719%
        z = 109,   1.2671%             l = 73,   0.8486%
        C = 107,   1.2439%             ` = 72,   0.8370%
        ) = 106,   1.2323%             F = 71,   0.8254%
        u = 106,   1.2323%             Z = 71,   0.8254%
        A = 105,   1.2206%             j = 70,   0.8138%
        & = 104,   1.2090%             h = 69,   0.8021%
        < = 103,   1.1974%             \ = 69,   0.8021%
        ? = 101,   1.1741%             k = 67,   0.7789%
        E = 101,   1.1741%             c = 67,   0.7789%
        / = 100,   1.1625%             V = 66,   0.7673%
        9 = 100,   1.1625%             T = 65,   0.7556%
        : = 97,    1.1276%             d = 64,   0.7440%
        $ = 96,    1.1160%             L = 64,   0.7440%
        y = 95,    1.1044%             O = 64,   0.7440%
        i = 94,    1.0928%             U = 63,   0.7324%
        1 = 92,    1.0695%             [ = 61,   0.7091%
        t = 92,    1.0695%             R = 60,   0.6975%
        @ = 92,    1.0695%             S = 59,   0.6859%
          = 92,    1.0695%             W = 56,   0.6510%
        q = 92,    1.0695%             P = 56,   0.6510%
        - = 91,    1.0579%             Y = 55,   0.6394%
        B = 91,    1.0579%             b = 55,   0.6394%
        I = 90,    1.0463%             _ = 54,   0.6278%
        7 = 90,    1.0463%             ] = 52,   0.6045%
        s = 90,    1.0463%

                               ###

V.  Errors and problems:


Problem:  Nothing happens, the computer just sits there.

You  have  forgotten to name the <input file.  As a  result,  the
program  is  correctly  (but stupidly) waiting for  you  to  type
characters on the keyboard.  (Same as <CON:)

Solution:  Type Ctrl-C or Ctrl-Break.
           This problem does not occur in Version 2.0.



Problem:  File not found.

You  have not spelled the name of your <input file correctly,  or
you have given an incorrect or incomplete path name.

Solution:  Check the directory for exact spelling.


Problem:  No result.

The first thing that should suggest itself is an empty data file.
As unlikely as this seems, it might be true.  Your directory will
indicate  0 bytes for the file size, if this is the case.   (When
you  see  the  evidence, you will remember  how  you  caused  the
problem  yourself sometime last week.)  A doctor I  know  created
this situation by using WordStar to split data into two files; he
then  DELETED all the records from the original file  instead  of
erasing  it.  It can take quite a while to diagnose this kind  of
error, so go "by the book" -- check your file size as a matter of
course.

Secondly, you might be asking for something that does not  exist.
Either  PICK  cannot  find an exact match in a  field,  or  FIELD
cannot  find  the 6th field in records that only have  5  fields,
etc.

Another  possibility,  you want PICK to find the  pattern 150 but
PICK  thinks 150 is a field number..!  Use PICK 0 HAS 150 instead
of PICK 150.  Also valid, PICK 0 150.

Solution:  Check your directory.  Check your spelling.
           Try to PICK using the HAS operator instead of EQ.


Problem:  Output contains garbage from a help screen.

You have run the output from a utility through a pipe into FIELD.
However, a command early in the chain bombed out and printed  its
help  message  into  the pipe.  When FIELD got it,  it  tried  to
format  the  output.   The result is bits and pieces  of  a  help
screen, possibly TSORT-ed!

Solution:  Put complex commands in a batch file.  Use FNKEY.



Problem:  No room on the disk

MS-DOS  pipes are actually FILES -- hidden,  system-level  files,
but  files nonetheless.  Every pipe you create will  demand  disk
space  until its job is done.  In addition, MS-DOS will  allocate
at  least  one entire CLUSTER to a file no matter how  small  the
file  actually  is in byte count.  Compare available  disk  space
before  and after deleting a very small file.  Are you  surprised
by  the result?  Erasing one small file will free up at least  8k
of  disk  space!  So a command chain that uses an input  file,  a
pipe, and an output file can demand up to THREE TIMES the size of
the original file by itself.  And the actual amount allocated  by
MS-DOS  for any fractional part of a cluster is 8k -- that is,  a
file  with  8,193  actual  bytes takes up  16k  of  space!   Your
reaction to this bit of news is probably the same as mine was.

Solution:  Erase some files and try again.  Get a hard disk.
           Get <input from Drive B, while logged onto Drive A.

                               ###