Re: GTK file selector

[David Kastrup]

Aidan Kehoe <address@hidden> writes:

>  > And yet, nobody apparently used this functionality for years and
>  > years.  A lot of the stuff in XEmacs is like that: implemented,
>  > and left unused because it has, maybe because of the roughness of
>  > APIs and documentation, not been tied into any application in
>  > frequent use.
>
> And maybe because anything implemented using the XEmacs-specific
> APIs will never run on GNU Emacs, so coders tend to put off writing
> code to use a feature until that feature makes it into GNU Emacs, at
> which point XEmacs provides a compatibility API--the reverse is
> never true. Eminently rational behaviour.

I can only speak for myself.  I tried using and working with images
when XEmacs was the only variant providing them.  I failed.  I was
completely unable to make heads or tails of the available
documentation.  The whole details with specifiers and instantiators
and stuff like that was completely opaque.  References in the manual
pointed you to other places in the manual, no examples were there,
terminology was not defined.  In short: it was completely unusable
unless you were willing to get intimate with the code itself and find
out what this was supposed to be all about.

The preview-latex project basically lay dormant until Emacs acquired
image support.  It became a matter of professional pride to make it
work under XEmacs, too, after it worked under Emacs already.  After I
had in the past given several rants about the quality of
documentation, things were supposed to have improved.  I tried several
times to get it to work, and failed.  Some volunteer tried his hand,
and gave up.  Finally I got a core developer from XEmacs interested in
the problem.  It took him months to port this to XEmacs, partly
because XEmacs image code was still not documented usefully for all
but the people really into the XEmacs code base, partly because the
XEmacs image code was chock full of bugs.  As well as the process
handling.  Bugs I could not possibly have imagined, circumvented, or
reliably tracked down.  Bugs in functionality that was claimed to have
been working for quite a few years.

preview-latex is an application that works with the purported strong
points of XEmacs.  It would have been utterly impossible to get it to
work without the help of an internal XEmacs programmer who worked out
how to do stuff that was missing in the documentation, and who fixed
most of the bugs that were simply everywhere.

In contrast, the documentation of Emacs was comprehensible, and stuff
basically worked.  There were redisplay errors: I think I sent at
least a dozen separate bug reports about them to Gerd Möllmann.  But
those were locatable, and not of the "things crash" or "impossible to
figure out how this is supposed to work" variety.  And the manual was
sufficient for finding out how things were supposed to work.

If things don't get ported to XEmacs until there is a compatibility
API, this can simply be because the native API is incomprehensible, if
it works at all.

Point the fingers all you want, but you are deluding yourself if you
imagine that the waning popularity of XEmacs among developers is just
a matter of religious beliefs.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum