Helmut Kopka's interpretation of the TDS

Michel LAVAUD twg-tds@mail.tug.org
Wed, 6 Nov 1996 11:54:11 +0100


To: twg-tds@tug.cs.umb.edu
Subject: Re: Helmut Kopka's interpretation of the TDS

> The TDS *is* voluntary, after all.  If someone decides to
> write/recommend stuff that is different from the TDS, and doesn't
> communicate with us about any questions, there's not much we can do.

Just to inform you (no will to raise any controversy):
In my new distribution of TeX & Co for PC, I have decided not to
follow TDS, because, after many weeks of thoughts and trials, I
came to the conclusion that PC users (I mean users of DOS / OS2 /
Windows3.x /95 /NT) could run into trouble very easily, in a way
they cannot detect nor predict, as soon as they create relatively
complex documents. And I would hate to provide arguments for
scientists to switch from (La)TeX to MS-Word.

The distribution is available by anonymous ftp on
ftp.univ-orleans.fr, directory /pub/tex/PC/AsTeX. 

I think TDS is adapted to Unix and not adapted to PC because of
(at least) two technical reasons, that are bound to each other:
1 - there is no equivalent of "ln" Unix command, neither in DOS,
nor OS/2 nor Windows 95 nor Windows NT. Therefore, users can have
several TDS hierarchies if they have several disks.
2 - Length allowed for environment variables is very short (127
characters in DOS and DOS emulators of OS/2 and Windows95/NT). 
This can put users into trouble as no OS warns when truncating
environment variables.

Other reasons are due not to TDS but to PC programs that have
been compiled with a different hierarchy. Using TDS is possible
but requires heavy use of environment variables, so we are back
to point 2. Increasing too much the environment space diminishes
space available for other programs in the 640K limit of DOS or
DOS emulators, and this may prevent people from running other
programs they need...

These are some of the reasons why I thought it was urgent not to
hurry to adopt TDS on PC. Of course, it is just an opinion.
Probably other implementors on PC have different ones.

Michel Lavaud  (lavaud@univ-orleans.fr)