*Note: I wrote this blog entry to accompany the IBM Data Magazine piece mentioned in the first paragraph, so for people following the link from there this goes into a little more detail on what RDF, triples, and SPARQL are than I normally would on this blog. I hope that readers already familiar with these standards will find the parts about doing the inferencing on a Hadoop cluster interesting.*

Expand those shortened URLs before archiving twitter messages

What if a shortening service goes down?

People love to talk about the implications of going down, but what if a URL-shortening service goes down? When I had trouble getting to recently, I realized that when they’re down tweets referencing URLs are worthless—and that it wouldn’t be too difficult to do something about it before this happens. (I have wondered, though: why doesn’t twitter grab some short domain name and offer their own shortening service?) After all, if you’re saving any…

My own little Twitter client

No AJAX, Flash, or AIR; just HTML, but arranged the way I want it.

I’ve tried various Twitter clients, but usually just went back to the web-based interface that people hate so much. My main complaint with it—and I saw no other clients that did any better—was that it showed tweets in reverse chronological order. Conversations and the multi-tweet mini-essays that some people write are difficult to read that way, so I decided to write my own little client.