There's one feature I really enjoy about Wordpress (I saw it first on Aaron's blog, but I can't seem to bring up any posts): related entries. Computers need to be able to mine data to find related content (I railed against tagging a long time ago as the result of failed tools).

I started hacking together something to try to find related entries - I managed to pass all 2000 of my entries through the related post filter, and I got some interesting results. What this feature does is it takes your entry's contents and compares it against your past entries' contents to see if they are similar. Similarity is handled by keyword and phrase overlap (nothing groundbreaking here). The feature took about 4 hours to implement from beginning to end, and I'm pretty happy with the end results.

One of my biggest regrets about Tabulas is not pushing for more of these types of discoverability tools - hopefully as I get some more time, I can hack together some more useful tools (along with a whole new concept of communities).

For example, see what the feature matched against my asstronaut entry. Pretty accurate, I must say! It's sort of fun to see what Tabulas thinks are "related" entries for a given entry.

Once I get this rolled out on a large scale to all patron accounts (sorry to those free accounts, there are technical limitations), I'll be interested to see how well it matches against my friend's entries.

Obviously this is a really broad first pass, so I'll tune the feature a bit to match some better things ... but it's a pretty successful first pass!