Let me please express my opinion and share experience.
I am not a long‐time groff user, but I use it exclusively
for about two years to typeset my papers and short booklets.
There are people here in the list who have used original
troff when I wasn’t born yet. Some of them participated in
the development of the Unix operating system and contributed
alot to it. I wasn’t going to babble in this thread, since
I am not active in the list and I do not consider my point
of view valuable to be worth of posting, but the topic
raised is close to my past and I want to share some thoughts
with the people here.
I tried a number of typesetting systems before I got
acquainted with groff. Many of them did the job well, LaTeX
is an amazing piece of software, no doubts. But none of
them gave me a feeling of typesetting. Probably, I am sim‐
plifying it too much, but isn’t typesetting is just a place‐
ment of glyphs, figures and special symbols at the right
places? I wanted to do typesetting, not markup, like I am
sitting in front of a letter board. With groff I can do it
in the way I like, if I feel that some glyph is not at the
right place, nobody can stop me from changing its position.
For example, using eqn I typeset a formula
.EQ
P sub g = 10
.EN
and it doesn’t looks nice, because ‘P’ has a lot of blank
space before the subscripted ‘g’. Manually I adjust the
space before the subscript
.EQ
P back 20 sub g = 10
.EN
and it looks more pleasing to my eye.
Someone before noted that troff is for people who are not
afraid of to soil their hands. That is what I need exactly,
I want to soil my hands and do the job well to please my eye
and probably someone’s else.
The supplementary tools like pic and eqn are also exception‐
ally nice. I used to typeset a small booklet on linear
algebra basics and found that typesetting formulas (even the
complex ones) can be easy and literate with eqn. When I
typesetted a short introductory paper on fuzzy logic, I
found that drawing control diagrams and blockschemes with
pic can be productive too. Besides, every aspect of both
formulas and pictures can be handtuned, if I want, and
nobody forces any policy on me. Yes, my hands were soiled.
One day I had to typeset a course material about machine
arithmetics. Standard Symbols font had no glyphs I needed
(circled division and circled multiplication). I wanted my
paper to be readable on any PostScript engine without any
additional fonts. I draw all of them, including circled
addition and circled subtraction, using groff escape
sequences and declared them as glyphs. Moreover, I was in
need to draw many bit boxes displaying internal representa‐
tion of floating and fixed point numbers. Hence I wrote a
couple of macroses to do this drawing job. Someone (Eric S.
Raymond?) once noted that groff’s markup is a humane one and
he was right. It is humane.
I can give many other examples where semi‐complex typeset‐
ting tasks were made easy with the help of groff. Surely,
all these things can be done in LaTeX too, no doubts. But
groff/troff ensures me that I can dig in the macropackage
and understand it (with a high probability), customize it or
write my own macro, if it doesn’t exists. Two very impor‐
tant things the groff/troff has, which seem to be problem‐
matic or difficult in LaTeX: it is able to produce plaintext
and move the baseline position and current location coordi‐
nates at any place of the page absolutely.
With the hehemony of LaTeX and word‐processing packages we
often forget the pedagogical aspects of groff/troff. But it
is another story...
--
Grisha
04.10.2012 23:09, Doug McIlroy пишет:
I had to chuckle at this
I can't believe it fits in 10M.
Back when troff was drilled into my fingers, it fit in 64K
program space plus 64K data space. Each preprocessor got
a similar allotment. Roff, the predecessor of troff, fit
in 8K total. Some features have been added since, but in
nothing like a ratio of 10,000:1 in utility.
Doug McIlroy
