Re: [Groff] troff(1) and info

[Bernd Warken]

On Thu, Nov 01, 2001 at 01:26:31AM +0100, Werner LEMBERG wrote:
> 
> As long as groff.texinfo isn't finished,

Documentation is never finished.  It's a work of a few minutes to move
the language parts of troff(1) into a new info section  "Differences...".
By and by tis information can be spread into the other info sections.
Man section 1 is for command line stuff only.

BTW, groff is a roff documentation system, so why use the TeX info system?
Man pages are better IMHO, and the groff HTML interface is powerful enough
to provide hyperlinked documentation.

> differences to Unix troff will stay in troff(1).  

So to use groff you always need the old-fashioned troff docu of the 1970's.
And you're fucqing-up the man page system, being the most famous 
application of groff.

> gtroff(1) isn't a good idea on platforms which
> use GNU troff as the default (e.g. Linux).
>
I can see 3 different possibilities.

1) groff is a stand-alone system that improves classical troff without
strict compatibility.  Then groff is not troff, but gtroff.  troff keeps
its own nature, parallel to groff.

In this case, we do not have a right to use troff(1) for documenting
gtroff.

2)  groff obsoletes troff.  In this case, using troff within groff is ok.
But then the groff documentation must be stand-alone, and may not be
defined in difference form.

3) groff is the hereditary of classical troff.  In this case, a strict
compatibility with the classical case must be offered.  Then the troff
manual would have bible character and the differential documentation
style would be justified.


Actually, groff is a foul mixture of all of these excluding cases.
By troff(1) and other things, it is not stand-alone, so 1) does not
apply.  As groff does not restrict to gtroff, 2) has gone as well.
And because even the compatibility mode is not compatible (we are
even proud about this) point 3) cannot be claimed either.

I know similar deadlocks from the Macrohard OS's.

Bernd