Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site phri.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!vax135!timeinc!phri!roy
From: roy@phri.UUCP (Roy Smith)
Newsgroups: net.unix-wizards
Subject: Re: a smarter soelim
Message-ID: <231@phri.UUCP>
Date: Tue, 14-May-85 13:11:46 EDT
Article-I.D.: phri.231
Posted: Tue May 14 13:11:46 1985
Date-Received: Thu, 16-May-85 02:18:54 EDT
References: <10446@brl-tgr.ARPA> <373@mtxinu.UUCP>
Organization: Public Health Research Inst. (NY, NY)
Lines: 51

> We gave up at Berkeley when confronted with this problem. [...]  So far as I
> know, there's no general solution short of running all of troff locally.

	This is only marginally related to the above topic, but here goes
anyway.  After working with various flavors of n/troff for the past 7 or
so years, I've come to the conclusion that we just plain need something
better.  A typical print command around here looks like:

   "soelim manuscript | bib -t pnas | tbl | neqn | nroff -me -Tnec-t | prt"

	Prt is a filter I wrote which pumps its standard input through a
crt's printer port.  I've added a new capability to terminfo, called
"slave".  I set TERM to "vt100-nec" which is exactly like a vt100 except
that is has the slave capability set to "nec"; prt looks up the slave
type and initializes the nec right after turning on printer controller
mode.

	Anyway, because there are 6 different processes going, you have
to stand on your head to talk to each one without confusing the others.
Using $ as the neqn delimiter is a pain because you can't easily use $'s
as arguments to macros you want to define.  Before we started using tbl a
lot, all of us here gradually shifted over to using "|" and that worked
fine.  But tbl generates "|" in it's output, so that's no good.  Also, if
you want to have citations in your bib data base with stuff like $ BETA
-galactosidase $ in the titles (we do), you have to make sure that your
neqn delimiter is "$" when neqn sees the bibliography.  So, people
learned to flip back and forth between various delimiters.  You can't put
double quotes in the bib data base because that confuses the bib macros,
so we have learned to use \(lq and \(rq.

	It goes on and on.  Nroff steadfastly refuses to underline greek
characters or punctuation unless you hack the terminal driver tables (but
genetic notation often requires it).  So, you learn to do horrible things
with local line motions.  Our poor secretaries have to learn how to
figure out how many backslashes you need to protect something with to
make it get interpreted at the right time.

	I'll admit that compared to the competition (no, I've never used
TEX, but from Knuth's book, it looks like it would be too tough to teach
non-computer types) nroff and all it's various pre/post processors are
pretty good, but I think we've really pushed n/troff to it's limit and
beyond.  I'm waiting for the day when we I can get some Suns in here so
we can use their "Macwrite with a higher education", but I don't see how
that will interface easily with bib (people around here have finally
learned to love bib, and won't want to give it up).

	So, the bottom line is, I wish there was something better, but
don't know what.
-- 
allegra!phri!roy (Roy Smith)
System Administrator, Public Health Research Institute