Mark has a slide deck on the whole complex of RDF/syntax/semantics issues that I thought was pretty darn right-on. His conclusion is pretty gloomy, that Semantic Web work is doomed until a few of these problems get sorted out.
Why Not http://data.example.com? · So far, I’ve been whining about RDF syntax and demanding the killer app, but haven’t really had anything to say by way of positive advice. So here goes.
In all these years, the most compelling Semantic Web vision
I’ve heard was courtesy of R.V. Guha,
and here it is.
Suppose someone mentions a company name, say Example Inc..
I am pretty confident that if I type
www.example.com into a
with a little bit of work I’m going to be able to dig out a bunch of
stuff about them, including their CEO’s name, their address, phone
number, and stock ticker, their financials, and blurbs about whatever it is
they make and sell.
And if I can’t, that’s going to make me suspicious about the
(Little-known fact: All domain names of the form
.org, and so on are reserved forever, just so you can give
examples of this sort without accidentally pointing at anything.)
So, Guha asked, why shouldn’t I be able to send a software agent to
http://data.example.com and pull all that stuff out without
having to poke around their idiosyncratic web pages?
This was in 1998, and I still wonder why I can’t do this, it seems
like a no-brainer.
Also, it seems like this is more or less exactly what they made RDF for;
someone invents a basic bunch of property names, and then any company can
invent their own special properties, and you need a little bit of
RDF-schema-like machinery to make new property names a little more useful.
Given something like that that developed a bit of viral spread, I’d use
it close enough to every day to give
RDF.net to whoever got it
And that would work pretty well, because
rdf.net would be a nice
base URI to root the names of your standard set of property names.
Just an idea.