Path: utzoo!attcan!uunet!littlei!omepd!iwarpj!pcm
From: pcm@iwarpj.intel.com (Phil C. Miller)
Newsgroups: comp.text
Subject: re: Why TeX?
Message-ID: <4020@omepd.UUCP>
Date: 1 Jan 89 22:42:25 GMT
Sender: news@omepd.UUCP
Reply-To: pcm@iwarpj.UUCP (Phil C. Miller)
Organization: Intel Corp., Hillsboro
Lines: 84

In article <318@imuse.uucp> ed@imuse.uucp (Ed Braaten) writes:


>What are the advantages of using TeX over nroff/troff? I've heard 
>a lot about it, but since I don't have access to it, I've never
>been able to try it out.  I chose nroff since it seems to be a 
>standard tool available under most implementations of U**X.  
>What are the major differences in the two systems?  Which is
>more reliable, portable?

Learning a text processing language is very much like learning a
programming language.  The process is as subject to pre-existing
personal bias, is as error-prone, can be as frustrating.

And before I express my opinions, I will quite plainly point out that
they are just that--opinions.  Programming languages, editors, and text
processing languages are all subject to personal biases, and I am only
offering my own personal biases and am not pretending they are facts.

First of all, to extend the programming language analogy, recall the
general statements people make about programming in higher level
languages:  the average programmer can produce a fairly constant number
of lines in one day.  The higher level languages pack a lot more power
per line.

[In my opinion,] The LaTeX system is very much like a higher level
language.  The TeX system is more like an assembly language.  *roff is
like microcode.  *roff can be a real pain when you're down to lining
things up in the final draft, trying to make copy that "looks good".
LaTeX is much easier to deal with because it has very intelligent
native behavior and not as much tweaking needs to be done in the final
draft.


>I had some trouble getting nroff to print all the special 
>characters I need (umlauts & Cyrillic) since the printer I
>have (Star NL-10) likes to see ASCII NUL's in some of its 
>control sequences.  It works now, but since I don't have 
>source to nroff, I had to do everything through the term
>driving table nroff uses (/usr/lib/term/*).  Fortunately,
>SCO provides source to a generic term driving table which
>you can use to roll your own.  How flexible is TeX in this
>area?  How portable would the solution be?

The *TeX programs produce device independent output files with the
extension .dvi.  A custom program is required to print a .dvi file on 
a specific printer, but there is such a printer program for virtually 
any printer you are likely to encounter.

The printer programs generally produce bit-mapped hard copy, making it
easier to print unusual things: mathematical formulae, non-english 
character fonts, etc.  *TeX usually cohabits with about a zillion 
font files.  The *roff programs seem to rely on fonts already loaded
in your laser printer, although I am really conjecturing here.  I have
had difficulty getting *roff to go beyond a small repertory of
tricks.

There are about a zillion style files rattling around the *TeX
community, too.  These style files correspond roughly to the macro
files in *roff.  For example, a university may have a style file for
their style of thesis.  I have seen style files for publishing articles
in publications of the ACM and other periodicals.

Finally, one of my favorite reasons for using *TeX is that nobody's
trying to make big bucks off of it.  The Gnu people distribute *TeX,
nominally for the cost of distribution, and much of the software is
supported by a large user community.  I like that; people GIVE AWAY
software which is actually useful and DON'T make you belly up for it.


>I would welcome any input from those of you familiar with both
>systems...
>
>Best wishes in 1989!



>Ed Braaten                    
>Intel Semiconductor GmbH     

Phil Miller
iWarp Program
Intel Development Operation
Hillsboro, Oregon, USA