A notch above a monkey

Usefulness of microformats

Another Wednesday, another talk. I was sorry not to see more visitors, but I guess we have ourselves to blame. Talk would certainly benefit from a better description, but it is a topic that I find interesting, yet hard to describe. The running joke of evening was that we should prefix all talk titles with PHP to lure all those developers who seem to respond mainly to known keywords. It’s a shame that curiosity is such a rare trait after a certain age.

By the way, the guy who invented OPML is Dave Winer . As fate would have it, he’s also father of RSS.

And it’s RSS that I was thinking about later on, when I was contemplating the role of microformats on web. First version of RSS came out some 8 years ago and it wasn’t until last year or two that feeds gained enough traction to actually become relevant. Still, most web users have no idea what they are and I believe really interesting uses of feeds are still to be seen.

I also believe the same is true for microformats. We are at the start of the road and we are yet to see really interesting stuff. I think there’s a definite need and opportunity to describe information on web in ways more suitable for further use.

Rip, mix and burn. Or something like that.

But I still don’t care about XFN .

Broken Apple

It turns out Apple introduced new iPod , iMac and iTunes too. Video watching is in. Now you can watch video on your iPod and you can buy (if you have access to iTunes USA store) episodes of ABC’s popular shows, but in ridiculously small format (320×240).

Personally, I doubt I’ll be buying anything made by Apple anytime soon. My iMac still hasn’t been repaired and I have no idea when it will be. I wouldn’t have spare macs lying around even if I could afford it and this makes them more or less unusable to me for professional work.

Dreaming about GIL

I’ve been thinking about Python’s global interpreter lock (GIL) again. Since I’m a glutton for punishment and haven’t been crucified in a week, you get to hear about it too.

Has anybody thought and wrote about avoiding GIL until you need it? What I mean is creating and using GIL only when you start using code that might not be thread-safe.

So, what’s wrong with this idea?

Things that come to my mind:

  • garbage collector is not thread-safe
  • majority of interesting modules are not thread-safe, so there would be little ROI
  • dynamic nature of Python means you can’t know in advance if you’ll need GIL and it would a bitch to fall back when you do

I guess it’s just a really stupid idea.