From: "Mike Champion" <>
>
> I just find it interesting that there seem to be two routes to the
> "semantic web" functionality, one via RDF/ontologies/metadata, and the
> other via the kinds of things that Google and Autonomy do. So far the
> latter approach seems to be yielding useful results much faster than the
> former, but it it's very interesting to think about how they could be used
> together.
Part of the reason that the Google/Autonomy route is more successfuly right
now is that it relies on heuristics to extract semantic bits out of a very,
very messy web. Heuristics can take you only so far. RDF et al., on the
other hand, requires that the web be clean to start with. But this requires
a lot of up-front work. Once heuristics hit their practical limits, Google
will only be able to go further if the web becomes a bit more clean. (After
all, my current "desktop filing system" can only take me so far. Eventually,
if I want to keep finding the things I need, I had better start organizing
this stuff). Conversely, strict adherence to RDF will only contribute to
the SW's lack of success. The RDF-based SW will have to loosen its necktie
and relax a bit if it's going to go further (or at least learn to live
side-by-side with the non-RDF world), just as Google will have to start
paying more attention to the more formalized SW metadata. And, as a result,
this sort of vulcan mind-meld of RDF and Google you speak of does look
likely. I would even go so far as to say that it's the only real
possbility...
---
Seairth Jacobs
