There has been a number of posts recently about XPath search queries in URLs in true RESTian glory, inspired by the Kimbro Staken's new Syncato blog software. Some highlights:

• Jon Udell's Kimbro Staken's XPath-searchable blog
• Sam Ruby's Introducing Syncato, Syncato thoughts, and Xpath enabled query
• Russell Beattie's Using RDBs and XML-DBs
• Mark Lussier's XPath plugin for blojsom demo
• also see Kimbro Staken's blogfor more on Syncato

I think this is all pretty cool, especially because I have gone back and forth a couple times about whether or not to use XPath in URLs as the means of expressing queries in the iCite net.

My main concern was that XPath was maybe too complicated for people to use, but seeing it in action in these various contexts gives me more enthusiasm. And, I still have the option of supporting ways other than XPath for forming queries.

However, one of the core features for iCites is that any query, once it is executed, is also given a "natural" URL (not unlike tinyURL) which then can be reused by others. So,

http://icite.net/icites/jayf//referers/a[contains (@href, '.icu')]

can become

http://icite.net/icites/jayf/referring_icites.html

The point of this is that query results are effectively re-published, either on the source site or on the querying site. And, then other people can query these results and aggregate them as well.

The other point of this is that queries can be formed by non-REST means, such as XML-RPC, and then made available through a (REST) URL.

The other thing to note is that iCites can transform their formats, for example, a query could request an XHTML version or an RSS version or something not XML, like RDF expressed in N3. This is a query that is run before XPath is applied, so for example, XPath can be applied to extract items from RSS rather than from XHTML.

I think I am going to get some coding done this weekend—I want to get rolling on all this again!