Imagine you're looking at a blog post. There's a button on the page that says "Find more like this". It goes off to a search engine that returns articles like this one from all over the web.

How would you code this? Can you do it now? You could just take the title and do a google search on that but it clearly doesn't work. It really needs a service where you drop the whole text in and it does some lexical analysis on it to produce a search query.

While we're at it, why is Google Blogsearch so bad? Compare a search for Ecademy in Google Blogsearch and Technorati.

I've been thinking a lot about Marc Canter's latest rant about APIs in social networking sites. Where is the conversation happening about the design of these APIs? FOAFnet is dead. The individual bits like ID, Microformats and so on each have their own mailing lists and wikis. So where can I go to find out the latest state of the art thinking?

And on RSS, Atom and Google. Google Base (remember that?) uses RSS and RSS extensions as one alternative to code up entries. But instead of just giving them a URL to a feed, you have to ftp a file to them which they then take up to a day to process. This is all bassackwards isn't it?


[ << What will happen to the recalled batteries? ] [ ProgrammableWeb: Web 2.0 Mashup Matrix >> ]
[ 26-Aug-06 8:22am ] [ ]