Phillip Pearson - web + electronics notes

tech notes and web hackery from a new zealander who was vaguely useful on the web back in 2002 (see: python community server, the blogging ecosystem, the new zealand coffee review, the internet topic exchange).


On autodetecting groups

Over the next few days I'm going to have a read of Raph Levien's work on network analysis and another look at the Google paper. It should be possible to use Raph's technique to find significant 'blobs' of people who share similar links. Nathan Newman suggested doing some sort of category/affiliation autodetection, and this is probably the way to do it.

One fairly basic idea would be to look for reciprocal links, and group together people with strong connections like this. If you link to somebody and they link back to you, you are probably having a conversation ... if you link to lots of people who all link back to you, you are probably part of the same community.

Has anybody out there done anything along these lines for blog connections? Get in touch if you have ...
... more like this: []

Second run: more blogs

I've run the ecosystem again with InstaPundit and VodkaPundit's blogrolls included; here it is. Now it should look more like TTLB's results.

I've sent my source code over to N.Z. Bear and am doing another run right now - as always, check here for the running totals.

Stay tuned for more improvements. I haven't done the thing that NZB has, and I'm sure my link detection isn't perfect yet. Let's see what happens ...
... more like this: []