Re: [Groff] [PATCH] groff_tmac(5): Discuss stripping macros.

[G. Branden Robinson]

At 2017-11-08T11:52:37+0000, Ralph Corderoy wrote:
> Hi Branden,
> 
> Generally, try and use few words whilst preserving meaning and avoiding
> ambiguity.  Folks have to wade through this when they're looking for an
> answer to their question.  :-)

Brevity is the soul of...something or other. ;-)

> > +In the
> > +.I groff
> > +source distribution, some of the macro packages (corresponding to
> > +.BR groff_hdtbl (@MAN7EXT@),
> > +.BR groff_mdoc (@MAN7EXT@),
> > +.BR groff_me (@MAN7EXT@),
> > +and
> > +.BR groff_mom (@MAN7EXT@)
> > +as of this writing)
> 
> "currently foo, bar, and xyzzy,".  Or don't add the dependency that
> needs maintaining?

Huh.  I'd make the "doesn't need maintaining" argument to support what I
already have, not your revision.

"Currently" can become wrong.  "As of this writing" acknowledges that
the statement may become inaccurate in the future.

> > +are maintained in \[lq]unstripped\[rq] forms that are then reduced for
> > +installation.
> 
> s/reduced/shrunk/?  Since that's the aim, and reduce has many computer
> meanings.

No objection.

> > +The stripping process primarily eliminates spaces between dots and
> > +request or macro names (used for indentation and readability of the
> > +macro sources) and most comments.
> 
> s/process //?

I would go you one better and get rid of the "The" in that case, too.

> (So it doesn't handle spaces after «'» too.)

Correct. '\" comments are rare and used to smuggle preprocessing
advice into man pages, though "%beginhere%" still serves as a guard
against that case.  It seems good to have an escape hatch for a comment
that for some reason really needs to survive shrinkage.

> > +If you strip your macros, we strongly recommend that you follow GNU
> > +.IR roff 's
> > +example
> 
> When did `GNU roff' get invented?

Well before _my_ time! :)

> I see groff(7) starts with
> 
>     The name groff stands for GNU roff
> 
> but I've always understood it was a contraction of `GNU troff', and roff
> was in no way a direct influence on its name or behaviour;  troff and
> nroff were copied.

The existing groff documentation is largely consistent in referring to
"roff" as the language, in contrast with nroff and troff as typesetting
programs--though I'll grant that the Texinfo manual that FSF policy
wants people reading in preference to all man pages exhibits a major
exception with its subtitle.

I'm attaching the output of "zgrep -C 1 -iw roff" of all the groff
1.22.3 man pages so that you can see where usge of the term stood before
I came along.  :)

It's definitely not my innovation, but it makes sense to me.

-- 
Regards,
Branden

uses_of_roff_in_groff_1.22.3.txt
Description: Text document

signature.asc
Description: PGP signature