Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site sdcarl.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!ittvax!dcdwest!sdcsvax!sdcc3!sdcarl!rusty
From: rusty@sdcarl.UUCP (rusty c. wright)
Newsgroups: net.text
Subject: Re: TeX on average crts.
Message-ID: <212@sdcarl.UUCP>
Date: Mon, 24-Jun-85 05:06:39 EDT
Article-I.D.: sdcarl.212
Posted: Mon Jun 24 05:06:39 1985
Date-Received: Wed, 26-Jun-85 06:29:24 EDT
References: <3285@dartvax.UUCP>
Reply-To: rusty@sdcarl.UUCP (Rusty Wright)
Distribution: net
Organization: Computer Audio Research Lab, UCSD, San Diego, Calif.
Lines: 62

>I believe that TeX doesn't have graphics capability only because TUG, or
>some other body, hasn't agreed on a standard yet. Certainly the program can
>handle it.
>
>Also, although it is no doubt possible to write a format to generate output
>on CRTs, as Rusty Wright suggests, it would perhaps be simpler to write a
>dvi driver. For example, on Unix* systems, this would allow considerably
>simpler access to screen routines for the output, and a dvi2screen program
>can handle font changes as well as a format.
>
>dartvax!karl                       karl@dartmouth.csnet

I don't think that it's that simple (to just write a dvi2tty
program).  You have to modify TeX as well (the plain.tex file).  For
example (the obvious one to me), if the user specifies using a font
that is 36 point (for example) then TeX is going to use both vertical
and horizontal spacing for 36 point letters while that font is in
effect (it derives this spacing from the .tfm file and the specified
magnification).  When the dvi2tty program prints the chars every one
in the file/document is going to be printed at the same size, about
10 point, (because they are on a crt, line printer, etc.) and the
spacing is going to look funny in those places where the user changed
the point size (magnification).  It just occured to me that there are
really two problems; changing fonts and changing the magnification.
In TeX you can get a different sized font by either loading that
font, for example cmr12 (computer modern roman, 12 point), or instead
you could load cmr6 magnified to 12 point (there is an issue of
design size that i'll bypass here).  You can also set a global
magnification that causes the entire document to be magnified (so
that you can subsequently photo reduce it).  I guess you'd have to
modify the plain.tex file to perhaps ignore magnification requests.
As i said in my previous note, i haven't really thought a lot about
this.  When i think about it i tend to think about what nroff does
when handed troff input.

I suspect that part of the reason that the TeX project isn't too hot
about the idea is because one of the nicest things about TeX is that
your output is always the same regardless of where you produced it;
you can run your input through TeX on an ibm pc, a vax running vms, a
vax running unix, an ibm mainframe, a 68000 based sun, etc. and
they'll all produce identical output.  Once you start using a TeX
that can produce output for a tty then you suddenly break that
uniformity.  I have personally been in the situation where we didn't
have a laser printer for many years and only had a diablo (daisy
wheel) printer and always used nroff.  When we got our laser printer
we had to go through all of the stuff we had produced with nroff and
fix it up so that it would look right with troff instead of just with
nroff.  When you only use nroff you inadvertently and unknowingly
wire into your document nroff specific stuff.  Even to this day i
have gotten documents that have parameters in them that are tuned to
that place's particular printer.  Also notice that troff/nroff
support conditionals that say "do this if we're troff" and "do this
if we're nroff".  And this is what i think the Stanford TeX group
doesn't want.  They want complete, homogenous uniformity.  This is
why there has been such venom from David Fuchs (one of the Stanford
TeX bigwigs) over Imagen distributing an old version of the Computer
Modern TeX fonts (Imagen claims that it is too much hassle to keep up
with the shifting sands of Computer Modern and when it solidifies
then they'll distribute that final version).
-- 
	rusty c. wright
	{ucbvax,ihnp4,akgua,hplabs,sdcsvax}!sdcarl!rusty