Re: Heirloom Troff for NetBSD (was: Removing ARCNET stuffs)

[David Holland <dholland-tech%netbsd.org@localhost>]

On Mon, Jun 01, 2015 at 02:11:03PM -0700, Greg A. Woods wrote:
 > > (quite seriously, I've been looking for a while for an alternative to
 > > groff for typesetting the miscellaneous articles in base.
 > 
 > I don't see how you can be serious, unless you've got objections to
 > viable alternatives that you've not yet voiced (so far as I can find).
 > 
 > Way back in 2009 on tech-userlevel I proposed the following:
 > 
 >   However now that I've looked more closely at the Heirloom Documentation
 >   Tools, I think they might be the best all-round replacement for GNU
 >   Troff in NetBSD:
 >   
 >   	http://heirloom.sourceforge.net/doctools.html
 >   
 >   They're already ported and running on NetBSD. They are much better
 >   supported than the DWB-3.3 release and already include UTF-8 support.
 >   They're possibly even better supported than the Plan 9 version.
 > 
 > You were reading that thread, and in your contribution to it you vaguely
 > suggested that your goal would be to replace troff input syntax.
 > However you did not respond to my proposal.

Knowing me, the mail's probably still sitting in here somewhere marked
for reply, with the intent of getting to it "soon".

Anyway, my thoughts on this subject are as follows:

 - we have a bunch of miscellaneous articles that we ship with base;
 - we should ship them as plain text, html, and pdf;
 - that we don't have html and pdf readers in base is irrelevant;
 - that roff is undesirable as a source form and I'm willing to
   translate the articles if that looks necessary or desirable.

Meanwhile, the criteria for a new tool are that it:

 - supports the needs of the articles we have (and reasonably
   foreseeable needs of articles that may appear in the future);
 - is suitable for import into base;
 - produces adequate and non-mangy text, html, and pdf output.

And I also extend this additional and somewhat more controversial
point:
 - we ought to be able to ship the NetBSD Guide as one of these
   articles.

The typesetting tool we have (an old groff) is inadequate because it
does not produce pdf and produces mangy html. It is also undesirable
because it's GPL'd C++, and because it's roff. (A newer groff does not
address any of these problems.)

The idea of adopting the original troff has come up many times, and as
far as I can recall the conclusion each time has been that it would
need quite a bit of work to be suitable, work that nobody seems to be
very enthusiastic about. If someone has now already done that work,
maybe it's become a viable option.

 > I also "vote" to kick out, or at least stop using, this mandoc thing.
 > It is a non-solution that more than muddies the waters.  Interesting
 > academically perhaps, but otherwise a major distraction.  Making it
 > equivalent to ditroff+pic+eqn+tbl+grap would be a waste of time and
 > energy and focus (though perhaps still academically interesting).

This is not sensible. mandoc is a substantial step forward.

 > Of course if _new_ significant documents are being written then perhaps
 > opportunity should be given to do so with some new tools.  My personal
 > vote would be for Lout.  It vastly improves on what troff can do, but it
 > also effectively replaces all the other associated and necessary tools
 > (pic, eqn, tbl, grap), and it generates PDFs as well.

Lout has come up repeatedly also. I don't know anything substantive
amount it, but for the most part nobody who does seems to think it's
particularly suitable.

 > It's no wonder NetBSD gets a reputation of being crotchety and
 > antiquated.

How dare we have criteria and goals for replacing things, even
substandard things? :-)

-- 
David A. Holland
dholland%netbsd.org@localhost