The Semantic Web, Part I: Promises, Promises

Over eight years ago I called the Resource Description Framework (RDF) "dead on arrival". I said back then:

The codification of the Resource Description Framework (RDF) in 1999 was eagerly awaited by those interested in describing electronic resources. Targeted to solve the problem of encoding metadata for digital objects using XML so that software could retrieve, parse, and index this information, it seemed like the weapon that would slay the jumbled mess that the web had become…

The weapon turned out to be a space-age laser. It could slay the monster if you could just understand it well enough to use it. Instead of holding to the rule of simplicity that is a hallmark of XML, as I hoped some two years ago (see "21st-Century Cataloging," LJ 4/15/98, p. 30ff.), a group of experts in database design and information retrieval (as part of the World Wide Web Consortium’s standards process) decided to build a structure based on directed labeled graphs. If you don’t understand the last three words of the previous sentence, neither will you understand RDF.

Unfortunately, this will kill it dead. If no one can understand how to properly use it, or if different individuals encode the same information differently (something that is already happening, even among RDF aficionados), then it will never fill the role it was designed to play. In the end, this was much too important an effort to be left to experts.

For those of you taking notes, RDF was to be an essential technical component that would enable the Semantic Web. According to no less of a source than Sir Tim Berners-Lee himself, "The Semantic Web will bring structure to the meaningful content of Web pages, creating an environment where software agents roaming from page to page can readily carry out sophisticated tasks for users." If you think this sounds like science fiction, you’re right. That (and more, much more) was written over seven years ago in Scientific American. How close are we to that? Well, look around. Where is your promised software agent that can scuttle off and perform "sophisticated tasks" for you? I sure don’t have one.

I think the problem is this — the vision, as compelling as it may be, is too far out in front of the reality. Sir Tim Berners-Lee and company have pegged the goal so far above our heads that folks like me who live in Practical Land can’t even see it from here. But this is pretty much where my dumping on the Semantic Web (at least in this two-part series) ends. I’m writing this because I’m actually changing my view on this, and mostly because I see efforts that have set the goal close enough that I can see it. I still don’t know if we’ll ever get to where my agent will take my doctor’s prescribed treatment, then scurry around the web and find providers in my insurance plan nearby who have good-to-excellent ratings and who can offer apppointment times during the holes in my schedule. I assure you, I’m not making this up.

But I’m beginning to really believe that there is something we can achieve with some of this technology that can be useful and worth doing. It is that which will comprise the second part of this series. Stay tuned, and meanwhile, what do you think about the Semantic Web? Can we do it? If so, how, and by when?