🚀 go-pugleaf

Being one of those souls who has been using (di)troff for a number
of years and refuses to switch to TeX (so far), I thought I'd share
my reasons, some of which have already been stated.

Although inertia is undoubtedly a major factor, I would be willing
to waste a few weekends learning a new system (e.g., TeX) and
converting any text I keep around into a new format; however:

  --  I move around quite a bit and, while I'm always (pretty)
      sure troff will work on the next Unix system I use, I have
      no idea if TeX will have been installed (or even if I could
      convince the appropriate party to install it).  Therefore,
      if I depend on TeX,  I can't count on being able to incorporate
      text which I bring with me (reports, memos, references, etc.)
      in a straightforward manner.
  --  I've (unfortunately) amassed some expertise in the nitty-gritty
      of troff formatting (especially ms macros), and I dread repeating
      the experience.  Since I'm not a `casual' document producer, I
      know I'll eventually need to make TeX do something for me that
      isn't straightforward.
  --  Not only does troff have nroff, but troff (ditroff, that is)
      can essentially use any font your driver/printer supports (when
      provided with the appropriate width tables).  TeX supports CM
      fonts, period.  Not only are they not widely available (and not
      all printers allow for downloadable fonts), neither are they of
      the same typographic quality as many fonts with which various
      printers are equipped.  See Knuth's comments on this subject.

Some of the complaints which have been voiced against troff have been
dealt with in various commercial (non-AT&T) troff products.  While all
of these require seperate purchase and installation (e.g., third-party
Documenter's Workbench products), at least it has been demonstrated that
troff can be just as `useful' a product as TeX (although, granted, it
does not do as well on mathematical formulae).

I don't want to open a can of worms over whose troff is better than whose.
Interested readers can take a look at "Geritol for Old Programs, or,
Troff's Got a Lot of Life in it Yet!", in the USENIX Winter Conference,
January 1985, pp. 165-169.  In there, you'll find discussions of some
modifications to ditroff which a few foolhardy souls in Toronto (myself
included) implemented, including:

  --  a hyphenation exception dictionary
  --  character kerning
  --  logical-to-physical font mapping
  --  font-table caching
  --  long font and macro names


Avi Naiman
Hewlett Packard & University of Toronto
avi%hpcehfe@hplabs.hp.com
...!hplabs!hpcea!hpcehfe!avi