[Groff] Choosing a portability target

[Eric S. Raymond]

I have been working hard at cleaning up the groff man pages so that
they use only constructions that are portable according to the third
draft of my report.  So far, this has been possible in all cases,
though I anticipate one -- groff_char(7) -- in which it will not be.

The groff project -- and I'm not sure what that means, other than
"Werner Lemburg" -- is going to have to decide how much portability it
wants.  There's a sort of sliding scale of difficulty here.

Target 1: doclifter compatibility

Easiest to get to from where we are right now would be doclifter
compatibility.  By the time I'm done my patching, we'll be there
for certain, because I'm using doclifter as a validator.

Target 2: man2html compatibility

Next most difficult, though not by much, would be to play well with the
C man2html used at the core of the KDE help browser. Gunnar was right,
this is clearly a close derivative of the code in the Brouwer man package.
I'm not yet certain, but I believe GNOME is using a minor variant of 
this engine as well.

The patching I'm doing now will get us very close to this, possibly
all the way there.  My goal is to get all the way there, but it may
take building some specialized validation logic and a second (much 
smaller) round of patches to the groff pages.

Target 3: manServer compatibility

This is almost the same as target 2 but for one single large difference;
manServer doesn't do conditionals.  All occurrences of .if would have
to drop out of macros visible to it.

Target 4: troff-classic compatibility

Gunnar Ritter has just informed me that if we want these pages to render
under AIX, Solaris, and HP/UX troff, we need to stop using long names.
All macro names would have to drop back to two characters.  (We would


The project will need to make a decision about the level of
compatibility to shoot for.  I can guarantee 1 without a lot more
work, but I am trying for 2 and believe we should consider that the
minimum acceptable goal.

I'm willing to clean up the pages to target 4 if so directed.  Offhand
I think 3 is probably impossible, but would be willing to do a careful
evaluation of the possibility if so directed.

Werner, this is probably your call.  Where do we want to land?
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>