bad at weaving?

I was reading the PyCon presentation at LJ (very interesting stuff) and got idly browsing the very cool SubEthaEdit notes (I will probably comment at some other time on my fascination with Python and about the enticing principles behind SubEthaEdit).

What I want to mention here is that one of the presentations at PyCon caught my eye: "GTK and Gnome" (on second day). As much as I hate to admit it, my pride started to itch. I would really like to be able to say wholeheartedly that I'm maximally open-minded with respect to all this gnome/kde stuff/flamewar/constructive-rivalry, but apparently I'm not. Well, I definitely rationalize about it, but my visceral/emotional reactions are quite low leveled.

Anyways, what got me thinking is that, hey, the KDE developers have a very good relation Python. And there are quite nice tools in KDE making use of Python or augmenting Python. But we didn't manage to publicize that at PyCon. And of course, this is bad (at least IMHO)! Lack of publicity! Neglect for the awareness the outer world has about our work/project/general-doings.

And then I tried to think about reasons. I'm far from having extracted a conclusion, but an observation is obvious: we're by far not as good as needed at larger-community weaving. I bet few of those living in the KDE/Python intersection realized that PyCon would be a nice place of publicity (healthily combined with learning and gaining conferencing experience). And of those, even fewer (none) had the time/energy/funding/volition to actually go there and do a presentation.

That's the observation. Now, the causes and the effects remain to be assessed.


...or traveling, if you're already in the USA. ;-)

PyCon would have been an interesting event to attend, but serious planning and preparation would have been required before making such an expensive and time consuming trip. Arguably, the recent UK event would have been as valuable to attend; certainly, the line up of "big names" from the Python world was most impressive.

There's also the question of what you present at this sort of event. The notes for the "GTK and Gnome" talk are quite interesting, though understandably brief. I'd be concerned about presenting features that people just take for granted.

Hopefully, PyKDE is about to enter the mainstream and get the sort of good publicity it deserves. Currently, the coverage any KDE-related solution gets on comp.lang.python tends to be limited, mainly due to the active role played by proponents of competing solutions, but there are also other familiar factors.

By davidb at Fri, 04/23/2004 - 10:15

but an observation is obvious: we’re by far not as good as needed at larger-community weaving.

I think you're right, I've personally done very little to publicise the KDE Korundum ruby bindings.

But I think one factor is that, although the Gnome people claim it's easy to write Gtk bindings, in fact it's very hard and they tend to involve teams of 10-15 people working for a year or more to produce something useful. I call this the 'knitting circle' approach - slow and sociable, so they have plenty of time and people to do the extra stuff.

Qt/KDE bindings are typically done by teams of 1-2 people, and are autogenerated to a much greater extent. The KDE ruby bindings were done by just Alex Kellett and myself, whereas the Gnome ruby project has 16-17 developers, and the api they've wrapped is about a tenth the size of Korundum.

By Richard Dale at Fri, 04/23/2004 - 11:04

Since the GTK talk notes mention Twisted integration, I should point out that Twisted has PyQt integration too. It works quite well, and I'm not just saying that because I wrote it. ;)

By taj at Fri, 04/23/2004 - 15:20

Reading about your "fascination with ... the enticing principles behind SubEthaEdit" gives me hope that someone starts to code something like this for KDE. (Given the cool kio_fuse and fish and KParts and other KDE technologies, this could become a realy ueber-cool thingie). Given that the 3 SubEthaEdit developers only needed 8 weeks to write it with Cocoa/OSX, it shouldn't take longer in Qt/KDE... Unfortunately I can't code. But ever since I saw it when 4 OS X guys hacked away together on a n ObjectiveC source file (each one seeing what the others were working on) and how productive they were, I was also fascinated by the idea.

Can you imagine groups of people clubbing together during Coding Marathon at aKademy, using "GroupKate" and jointly koding kool KDE stuff?

Your remark about KDE being "not as good as needed at larger-community weaving" is spot-on. We are not bad -- but we need to become better. We can change that with big events like aKademy and even more with smallish contributions and activities by each one involved in KDE...

By Kurt Pf. at Fri, 04/23/2004 - 21:57