A few days ago I wrote
about setting up a blog aggregator. Since then, much water has flown under the bridges.
I bought a domain: feedworld.net. The theory of computing aggregator is now at feedworld.net/toc
I set up a couple more aggregators just for fun, because it's so easy: Seattle startup scene
and Staring into the singularity
Behind the scenes, I upgraded the software to something called planet venus
. There's something else called plagger
that's supposed to be really powerful but is horrendously difficult to install; I gave up after several hours of frustration. Anyway, venus is powerful enough and lets me do a bunch of things:
I've added two new feeds
: papers uploaded to the ECCC and the relevant categories on arXiv. ECCC produces badly broken RSS feeds; I wrote a venus plugin to fix it.
In addition, the plugin/filter also has features to link author names to DBLP, automagically produce a download link and so on.
Finally, I added face pics for the people for whom I could find pics. If anyone objects to their picture I'll remove it.
In short, the aggregator is now at a point where it's vastly more useful
than subscribing to the individual feeds. Please give it a spin
Finally, there's someone else who's working on something called FeedWorld
but I didn't know about when I bought the domain; they were kind enough to let me use the name. I figured the least I could do was throw them a link. Thanks dude!
In the course of the last few days I've learnt a lot about RSS. It's got a lot of potential to be the "Unix pipe of the Internet", as many have hoped for. However, before that can happen, two things need to change: there are a zillion mutually incompatible formats; I hope eventually RSS 1.x and RSS 2.x will die out and Atom will win. Secondly, parsers need to become error-tolerant just like browsers. Parsing data coming from the web using a general XML parser is a pain in the butt.