M Bianchi <[EMAIL PROTECTED]> wrote:
> On Mon, Sep 17, 2007 at 01:52:04AM +0200, Gunnar Ritter wrote:
> > :
> > All variants of the original -mm which I have seen so far
> > have been very similar. It is likely that they all use the
> > :p register in the same way.
>
> One definition of backward compatibility is "All Bugs Are Preserved".
Err - but where is the bug here?
> So I ask the question of the group:
>
> Do we want to implement "backward compatibility" of undocumented
> things like the number register :p in the groff package?
>
> I vote no.
I personally do not care. I will always keep "my" variant
of -mm such that it retains compatibility with existing -mm
documents (regardless of whether they had been documented
in a paper that has been inaccessible to most people for at
least a decade). Since I generally intend Heirloom troff to
remain compatible with groff, this also means that anybody
who has a -mm document but wants to use groff should be able
to use my -mm variant as well.
I also do not care much about new features in -mm or the
like; in my experience, one is much better off to write
new in-house macro sets than to fiddle with the various
inadequacies of the old macro sets anyway. In fact, I
think that the strength of troff is much more the ease
of writing new macro sets than the existence of old ones,
all of which turn out to be quite limited once one wants
to do something not envisaged by their creators.
So for me, a partially compatible but extended -mm variant
would be useless. But your mileage might vary, and since
the original -mm is Open Source anyway now, it might be a
reasonable direction for a reimplementation to do something
different.
Gunnar
