> Very nice, dankeschoen. The problem was, I didn't know of eqn
> magic sdefine, as groff docs are in many places incomplete.
> \[de] will do the job, yes.
Well, "sdefine" is clearly described in groff's eqn manpage --
it works like "define".
Anyhow, I guess most of groff's documentation expects you to be
familiar with troff (just like the manual page for a C compiler
will describe the peculiarities of that particular C compiler,
but it won't teach you how to program in C). Everybody
working with groff should have read the original Bell Labs
documentation, in particular the Troff User's Manual (cstr#54)
and the papers describing eqn ("Typesetting Mathematics User's
Guide", 74/eqn) and tbl ("Tbl -- A Program to Format Tables",
76/tbl) available from http://cm.bell-labs.com/cm/cs/cstr.html
and http://cm.bell-labs.com/cm/cs/papers.html.
> Still better than LaTeX horror.
Do you mean horror with regard to function composition
operators, or horror with regard to LaTeX per se? If it's
the former, then I don't quite follow you, it looks pretty
much like it is supposed to:
\documentclass{article}
\usepackage{txfonts}
\begin{document}
$(f \circ g)(x) = f(g(x))$ alignment test: $+ \circ + \circ +$
\end{document}
If it's the latter, I don't quite follow you either -- LaTeX is
a wonderful program that works very well for the most part and
has loads of add-on packages and excellent support for PDF and
cross-references and hyperlinks. There's a certain reason it
has been adopted by many journals heavy on math -- there's no
substitute for quality typesetting.
(Of course, there's also a certain reason for some journals
wanting manuscripts in Word-format, even if it's only that
they don't expect their authors to get it right anyway,
and resign to re-typesetting everything submitted to them.)
