Sam Ruby is demonstrating, in his Data Flow post, the creation of custom XML "feeds" though using XPath queries in his site's URLs. I think this is awesome!

This is a good example for what I am building for the iCite net—it is like the custom feed "flavors" that this blog has via blojsom combined with XPath queries to customize the content.

One thing additional that I am looking to do with the iCite net is add the ability for any feed to combine feeds from multiple sites, and also things to have globally permanent and unique ids, so that if they are retrieved simultaneously from multiple locations, they can still be identified as being the same content. So, the idea is that iCites know that other iCites offer custom feeds, and can negotiate with each other for the desired content.

Even with Sam's example, imagine if your feed reader *knew* it could negotiate feeds with Sam's weblog. Your reader could *know* your preferences, and grab and process feed data based on what you are looking for. It wouldn't matter if the feed was RSS or even XML, as long as the reader could find out if the site offered the format it wanted.

(Part of the idea with the iCite net as well is that iCites are feed readers, and feed readers can be iCites, i.e., not being limited to the current client/server model.)

P.S. I posted on this topic recently in XPath URLs and web searches.