At ETech (which I unfortunately could not attend) there was a presentation entitled real world semantics that is close in spirit to my own recent experimentation. The presenters were Technorati's Kevin Marks and Tantek Celik, who fought the good fight to bring quality CSS support to Microsoft's now-abandoned MSIE/Mac. Phrases they use to define real world semantics: "emerging semantic (x)html", "adoption by 'real people'", "beyond academics and theoretical discussions." Exactly.
Meanwhile, over on Kingsley Idehen's blog, you can see another implementation of the kind of xhtml-aware search technology I've been playing with here. The advanced search feature uses the Virtuoso engine to perform not only XPath search, as I'm doing, but also XQuery search. Here's one of the provided examples:
for $i in node()//a return <p>{ string($i/@href) }</p>
This query, which finds links and produces a series of paragraphs containing the referenced URLs, shows how XQuery can combine the search capability of XPath with the transformative and generative power of XSLT.
Although random XHTML can be mined more fruitfully than you might suspect, I'm on the lookout for ways to naturally, and virally, enrich its semantic carrying capacity. The Celik/Marks presentation points to several such efforts, including GeoURL, which I use in my blog's header to announce my location (<META name="ICBM" content="42.93564,-72.27239">), and XFN, the XHTML Friends Network, which proposes using the REL attribute of links (<a href="..." rel="acquantaince">) to indicate relationships. This is the sort of thing that will make the search techniques Kingsley and I are demonstrating come alive. My hunch is that lots of XFN-like strategies will emerge, if we can close the feedback loop and connect the effort required to adopt such a strategy to an immediate reward.
Former URL: http://weblog.infoworld.com/udell/2004/02/18.html#a919