September 19, 2006

Microformats and URL Pipelines

Just before I went on holiday, I was having an email/GTalk chat with Brian Suda about various things and he showed me the neatest pipelined URL that I've seen so far, recalled here from the wonderful chat archive thing(?!) that chat in Gmail offers.

Brian Suda's crazy URL pipeline, escaped umpteen times....

The pipeline starts with http://suda.co.uk/publications/EuroOSCON06/ which deails the conference presentation Brian was demoing this sort of technique.

That post is then translated into German using the Altavista babelfish service (err...why - just showing off?!;-)

The German version of the page is then tidied up using the W3 HTML Tidy service to ensure it is well formed XHTML.

This document is then parsed by a routine of Brian's that makes a KML file from the scraped geotag info.

Finally, the KMl file is piped into a google map.

Phew...! One of the big tricks I picked up from this is the use of multiple escapes (which I realised is a trick that solves quite a few problems I've been having with deliSearch handling of URLs that need escaping...:-)

Anyway, apart from posting that megalink, what I'm really posting about is Brian's new O'Reilly PDF book on microformats... one to get added to the OU SafariU account, I think...

Microformats will have a lot of traction over the next 18 months or so, I think, and for all their downsides, I think they'll help the ad hoc semantic web reach some sort of critical mass. So much so, I think we'll see enough of an uptake in microformats to prompt a flurry of services in the same way that there are all manner of feed based services exploiting a rich enough feed ecosystem for them to be (almost...) viable.

Posted by ajh59 at September 19, 2006 12:12 AM
Comments