Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: Re: Questions About SUO

Seth,

It's an interesting question why there isn't yet widespread adoption of RDF in Web development.  (see for instance http://www.xml.com/pub/a/2002/05/15/rdfprimer.html for a different view on this).  Certainly I won't argue that RDF is perfect. But it's lackluster acceptance among the Web development community thus far might be due to other factors then technical snafus.  In particular, one could question whether RDF lies a bit beyond the immediate structure of incentives for developers, such that its use is still viewed as non-critical and superfluous.  For that matter, it seems that there is a sort of horizon for technologies involving Web development: inside that horizon you get adoption, on the outskirts things are "fringe".  An assumption that AI-enthusiasts often make is that ! the horizon is constantly expanding outward, and their tools and ideas will one day reign.  In that case, it is just a matter of time before RDF is adopted on a broader scale, and the next chink in the grand armor of the Semantic Web (or, a fortiori, the KR project) vision is in place, etc etc.  But that vision is based on some assumptions that these "fringe" technologies are really valuable and useful, and will revolutionize everything once the proper conditions are in place. I for one hope that is true.  But my lead-in here to the SUO focus is that we'd better start asking ourselves hard questions about what exactly an ontology is supposed to do (which will in turn guide the way it is developed), and what it is not.  In particular, it seems absolutely pertinent for those who wish to see SOMETHING by way of an improvement on search and retrieval, data reusability and app interoperability to be able to com! municate about what needs to be in place.  Also, about what already IS in place and how that might be used.  I agree with Todd (offline discussion) that something like a "killer app" that is wed to some particular ontology will probably be necessarily to pull naturally diverging (err, academic) strands of thought together and push everything forward.  Of course, the killer app itself will require that much of the ontology development already be in place....

Erik

http://www.xml.com/pub/a/2002/05/15/rdfprimer.html

  Seth Russell <seth@robustai.net> wrote:


From: "Todd Hughes"

> Thanks for weighing in on this issue. I'm inclined to agree with both you
> and Bill Andersen. In the context of something like HPKB, it's quite true
> that you can use a standard ontology to integrate several applications.
> However, in the context of the Semantic Web, you're talking about
> integrating web pages, not applications.

Huh ???? What gave you (and others) the idea that the semantic web, which
will probably end up being more like a data cloud than a web, must just be
about web pages? Imho, that is mis representing not only what is currently
being deployed, but also what it can achieve. If fact, describing web pages
in RDF does not even seem to be happening ... probably because somebody
screwed up how RDF integrates with HTML.

Seth Russell


Do You Yahoo!?
LAUNCH - Your Yahoo! Music Experience