Re: That old flamewar again [was Re: recent smiley.el changes]

[Dave Love <d.love@dl.ac.uk>]

Jan Vroonhof <jan@xemacs.org> writes:

> I hope the readers of 'ding' who must be all to
> familiar with this particular flamewar will forgive me this message.

I don't know what `this particular flamewar' is.  I'm apparently being
told off out-of-the-blue for doing something useful (well, smiley.el
was mostly useful just as a test, and has wasted too much time) and
then being taken to task for explaining the background.

> > Note in this connexion that writing common image code is messy since
> > XEmacs maintainers didn't cooperate on a common API 
> 
> This lies somewhere in between slander and 'the pot calling the
> kettle black'.

Please reconsider.  I reported simply what happened in that case and I
believe I've been cooperative; at least you've thanked me for it.
That's surely what you'd expect of someone working for GNU.

> Work on Emacs 21 started (or appeared to) long after I joined
> the XEmacs development and I don't recall _any_ request for such
> cooperation.

Presumably you simply weren't informed of it by who was approached.
You can't conclude an approach wasn't made (and, indeed, followed up).
I guess there's more background people are ignorant of.

> At the same time request for access to the fledgling Emacs 21 for
> API comparisons from interesting XEmacs maintainers were being
> obstructed.

I don't know what that refers to, but I won't conclude it didn't
happen.  However, I was introduced to the new redisplay by an XEmacs
maintainer who contributed to the API and I doubt anyone was turned
away who offered to do something useful for Emacs.  Note I said
`_co_operate'.

> The only inter-operability work was actually done by poor elisp
> developers being hurt by this. Per and Bill, you know who you are
> :-) and Thank you.

That's hardly fair to Gerd and others.  Note that I worked not only on
XEmacs-isms in widget/Custom, Gnus and W3, but also introducing XEmacs
features into the Emacs core.  The logs tell part of the story.

> Luckily the latter has been solved, let's hope it is not too late.[1]

If that means there basically aren't problems with interoperability,
or that, for instance, W3 is yet clean enough to include in Emacs,
it's optimistic.

> Maybe it is not enough, but other people reimpleting APIs have made
> do with less.

There was no wish to re-implement the XEmacs API, just (for me) to
understand what code was doing.  Emacs 21 compatibility with XEmacs
was always the aim, all else being equal, but the redisplay API was
partly a reaction to the XEmacs one.

> I am fairly sure XEmacs was already using CVS when smiley.el was
> created and Lars has been making ChangeLogs since day one[2]. 

It's not a question of whether things are under CVS or whether logs
exist, it's whether they allow verifying authorship.  I speak from
experience, and I doubt anyone else would have been significantly
better at it.  If Gerd and I hadn't spent ages on this stuff, chased
papers and made sure some code was re-written, Gnus 5.9 wouldn't be in
Emacs 21.  If the provenance of XEmacs code was generally clear, more
would probably have got imported to Emacs (by me, if no-one else).

> and don't use it to vent some kind of (unbased IMO) grudge.

I'm presenting relevant facts.  If people don't understand, it won't
change.  What I said is directly relevant to Gnus and the non-trivial
amount of work I've done on it.  Much of that was unfortunately of a
wasteful nature, but someone had to do it in the circumstances.

> P.S. I will NOT reposond to follow-ups on this message on
> non-appropriate fora, such as the ding list, if at all.

I've no idea why it would be inappropriate to explain maintenance
issues concerning a GNU project on a list (I don't currently read)
devoted to it.

-- 
We want to cooperate, but we are not doormats.  -- rms