Just as I sit down to actually blog something, I see the first gigantic spider of the summer on my wall, and have to run away screaming like a little girl. The universe doesn’t want me to write! (Let’s pretend every other week for the past six was stalled by way more spiders.)

• I fixed excerpts on ye olde blog, so more than a sentence appears on the front page. Hurrah.

• veekun has consumed most of my time, and probably put me in a horrible slump for a while. I’ve crawled back out of my hole at last, again, and am delighted to be excited about building things.

  • The Pokédex database schema was rewritten to support spitting out pages in any supported language, and I got all nitpicky about the way it was done and decided to fix it up “real quick”. Cue two weeks of mucking about with SQLAlchemy wizardry. At the same time I was stubbornly refusing to learn that I cannot, in fact, get much work of note done while chilling with my tiny laptop on Mel’s bed watching her draw. This has been a hard-won dual victory.

  • A lot of crap broke because of this change. I fixed most of it, maybe.

  • Black and White items grew some effects, and I rewrote some of the old effects (with syntax that no longer works).

  • And honestly that was most of it. Some little things were fixed or improved, and there was some whole mess with move metadata, but not much user-facing of note.

  I intend to break the media (200+ MB of ripped sprites, official art, etc.) out of the repository tonight, and consolidate the proliferation of littler repositories for various parts of the site. This will break everything again, but it’s been a long time coming and will be an improvement overall. Then I’ll finally have all the major bottlenecks out of the way, and I can start working on the lengthy list of interesting ideas I’ve accumulated over the past month.

• Less activity on floof, which is unfortunate, but I’m trying to give it some TLC lately. I did finish the whole art browsing thing, and now I’m trying to make MogileFS play nicely with Python. I read a few articles targeted at startups over the past couple weeks, and despite my meticulous launch plan, I’m convinced that I should just get the thing working and “launch” it. Maybe not advertise much for a while, but I figure a half-working site with some users is better than no site at all.

• I read a lot of stuff about Twisted, and I have a decent grasp of how it works. dywypi 2.0 is thus coming along decently, and shouldn’t take too much more work before it can replace the dywypi currently sitting in my channel.

Both at work and while working on dywypi, I’ve learned that I am really really bad at architecture design. I resist actually writing anything until I have a decent idea of how it works, but I don’t really have a good grasp of how something works until it’s being used, so I end up twiddling my thumbs and jotting down endless notes for far too long. And then what I actually implement ends up being inappropriate anyway.

The only cure seems to be acknowledging that this is a problem, being willing to refactor, and writing something that works before all else. Premature generalization wastes as much time as premature optimization.