Re: To XML or not to XML?

[Peter Simons]

Bastiaan, I am sorry for the long delay!

You wrote:

 > I am the author of the bnv_have_qt macro, which is one of
 > the very few that have been converted to XML. Although it
 > looks nice, I understand from a private conversation with
 > Guido Draheim that the XML initiative has been abandoned.

Well, yes and no. It hasn't been abandoned per se, but
nothing has happened in that direction for the last 12
months or so, so you could say it's dying slowly.

The problem XML was supposed to solve is not just that we
need cooler layout for the documentation -- it is that
macros need way more meta-data than they have right now. We
need to specify which license it's under, we need
dependencies between macros, packages of macros, we need to
mark obsolete macros, and so on and so forth.

Adding all this information into the macro in a clunky
syntax like @license or @depends-on is one way to do it, but
already 90% of the time I spend maintaining the archive goes
into fixing the use of these keywords! We have only a
handful of them, yet I usually have to edit a macro quite a
bit when it's submitted for the first time. (Sometimes I
have to do these edits every time the author sends in an
update!)

So when I chose to go towards XML, my hope was that
automated syntax checking and the wide availability of XML
editing tools (Emacs' PSGML rocks, for instance) would cut
down the time I need to spend on fixing the submissions!

Now, we already get a significant number of submissions
which ultimately don't make it into the archive because the
old mark-up is not powerful enough to present them
appropriately. Similarly, we have macros which clearly are
obsolete and need to be marked so. All these things are real
life problems that need to be solved, and if they aren't,
the archive will, well, not prosper. :-)

So to answer your question, Bastiaan, going back to the old
markup strikes me as the entirely wrong thing to do. I'd
much rather go _forward_ with getting more macros into the
new format.


 > The problem with today's situation is an inconsistency
 > between this archive and Guido's archive on SourceForge.

Yes, that is very annoying.

I am not sure what to say to this problem, though. I have
always hoped that the need for having two archives in the
first place would go away, eventually; that a better and
more powerful system would replace both of them. A new and
improved macro mark-up format was a key part of that idea.

I'd be interested to hear what other people (on the list)
think. Do you think it's worth to try and get the whole
system up and running in the improved state? The software
can do it all already, all that's missing is the effort to
convert the existing macros (and to clean up the whole
repository, while one is at it). Is anyone willing to help
with that?

Or do you have entirely different suggestions how to
proceed? Guido, what do you think?

Peter