Sporkmonger

If you can see this, apparently everything is working.

Update:

Ok, so there were some bugs. Incorrect permissions and such, not-quite-right Apache settings. But it’s likely to be more functional now.

After several weeks of work on the site, I’ve finally gotten everything moved off Typo, redesigned, and generally just looking a thousand times better than before. I also hosed out all of the accumulated spam. It wasn’t easy.

At Aristotle’s request, I made sure to modify the Typo conversion script (as well as Mephisto itself) to allow the feed entry GUIDs to remain the same. I wasn’t particularly pleased to discover that Mephisto doesn’t actually store its entry GUIDs in the database—it generates them each time the feed is requested, which in my humble opinion is a terrible idea, since the information it uses to generate the IDs might change. The change I made added a “guid” column to the database and uses UUIDTools to generate unique ids instead.

I also deleted a bunch of entries that I no longer felt were relevant to the site. I know a lot of people in the pro-standards crowd don’t like it when people do this, but I did anyways. I also tried to set up redirects for all of the old resources to the equivalent URIs in Mephisto, but I’m sure I missed a couple. If you bump into a 404 that you think is a mistake on my part, I’ll try to get it fixed as quickly as possible.

In addition to the switch to Mephisto, obviously I also took the time to redesign the site. I figured I should construct the design entirely from scratch this time instead of simply altering the default Typo theme. I gotta say, I really, really, really like the new look. I haven’t bothered testing and seeing how it looks in Internet Explorer. I really just don’t care that much about IE anymore. But I didn’t do anything particularly tricky with the design either, so it’ll probably look mostly OK. If you’re reading this in IE, stop, go install some other browser (preferably Firefox), and then come back. Web developers everywhere will thank you for it. And you’ll feel better about yourself. I promise.

I released a new library the other day called Squish. Admittedly, in its current state, it’s really effectively useless. I’ve never been particularly good at optimizing for performance, and while I did manage to improve the speed of the library (with the help of people in #ruby-lang) by at least an order of magnitude, it’s still way too slow to be useful. In retrospect, I’m not really sure why I even tried to write it in Ruby—I really should have known that a compression-based text classifier would need to be written in C. But hey, at least I know the basic idea works now.

Oh, and you may also notice that the site is faster and less error-prone now. You can thank Slicehost for that. I can’t recommend them highly enough if you’re in the market for a Rails-friendly VPS host.