Tuesday, 4 February 2003

Pete Cole responds to Don Park: Finding Hub Blogs. Blogspace needs a directory of hub blogs organized by topics.  Recognizing a hub blog is easy enough.  A hub blog has a large number of subscribers and links to other blogs.  A hub blog's topics can be either inferenced by content and links or specified explicitly. [Don Park's Blog] Maybe I've missed something but it does surpise me that the blog crawlers (e.g. Technorati) out there don't seem to do anything like this. The 'citation analysis' algorithm proposed by Jon Klienberg would be highly usable here, though would require a fair number of tweaks to sort out outbound links and inbound links to a 'blog' rather than an individual blog page - but results from this algorithm can be impressive. This is a cool idea - something I haven't even considered doing for the ecosystem, but it would be very sensible.  If anyone wants to have a crack at it (all the ecosystem data is available for research etc), do get in touch and tell me how you went. 9:58:54 AM    comment []

Sender Traumwind: Cherokee is a tiny, ultrafast, lightweight Web server. It is implemented entirely in C, and has no dependencies beyond a standard C library. It provides only the most basic HTTP functionality, but is extremely fast and small. 9:56:29 AM    comment []