Automatic Link Discovery
I've been deep into experimental code, backing an automatic link discovery mechanism based on my old Link Apprentice into Tinderbox, chiefly to see how it works as a weblog tool.
Think about it: a lot of us have a ton of writing in our weblog archives. Finding a way to dig it out would be nice. Tools like "a year ago today" are a help, Fagerjordian links are a help, but most archives pretty much just sit there and wait for Google.
One catch is that it's hard to know when the algorithm is being very clever, and when it's simply broken. Right now, it's suggesting a few links for this note:
- Category Factory (another attack on the same problem)
- Scallops (part one of this post)
- Airport Security Silliness (I have no idea what it's thinking)
Similarity is, in general, a greatly overrated reason to link. But weblogs need more links; it's worth an experiment.