Re: .TQ to replace .PD 0

[Ingo Schwarze]

Hi Ralph,

Ralph Corderoy wrote on Mon, May 23, 2022 at 10:22:58AM +0100:
> Doug wrote:

>> I fail to see any case for deprecating .PD.

> It seems pointless for GNU Groff to attempt to deprecate .PD when it is
> only one of the man-page formatters and has no control over the many
> existing man pages.

We actually do have partial control over a significant portion of existing
manual pages, by virtue of some relevant people participating in the
list <groff@gnu.org>.

I am (second to jmc@) co-maintainer of the manual pages in the OpenBSD
base system, and i regularly have contact to manual page maintainers
in FreeBSD, NetBSD, and (rarely, because they appear much less active)
DragonFly BSD and Illumos.  I do occasionally provide advice to
other BSD manual page maintainers, including questions of style,
portability, deprecation and the like.

Leading developers from the Linux Documentation Project also
regularly participate in this list, so we have almost all practically
relevant free operating systems covered to some degree.

Admittedly and for good reasons, huge numbers of individual, portable
software packages also provide manual pages, and we have no direct
access to the developers of those.  Then again, when those people
actively look for advice, they are most likely to look at what we
produce, in particular manual pages and style guides included in the
groff distribution, in the Linux Documentation project, and in the BSD
base systems.

So i think that in general, it does make sense for groff to provide
style advice and formalize deprecations, even if the time for such
advice to spread probably needs to be measured in decades and certainly
not in months.


Regarding the specific question of deprecating .PD, i suspect it
may be better not to, because it causes little harm and has some
applications where it is more portable and not significantly more
ugly than the alternatives.  But i don't feel strongly about that.

Regarding new text formatters and markup languages, i don't see
much need for them.  Over the decades, most other attempts turned
out much worse than ROFF and LaTeX, including practically all that
are significantly younger, and for software documentation in
particular, ROFF remains vastly superior to TeX and to all other,
younger solutions i'm aware of.

So i think carefully maintaining and slowly evolving the roff-based
languages is actually more promising than hoping for Go & friends.
Of course, compatibility should only be broken when that is unavoidable
for *very* important new functionality.

Yours,
  Ingo