[
Lists Home |
Date Index |
Thread Index
]
>On Wed, 23 Jul 2003 21:53:58 -0700 Tim Bray <tbray@textuality.com> wrote.
>
>OK, I have a question. How would you, given the existing deployed Web
>technology, be able to distinguish which of these two stories are true?
>
>I don't think you can.
Of course the deployed web technology isn't sufficient. We *know* that.
That's not the issue. The technology needed to make the semweb viable
requires us to be pernickity and very precise in what we tell our software
(or more accurately, what statements we write down for them about the
world). The issue is that the current set of web specs are not sufficiently
pernickity. That is simply a fact - no amount of argument changes it. It's
very simple. If we want a semweb, we must tighten up the language. If we
don't want to tighten up the language we don't get a semweb. If the
formalist can't make sesne of the specs for their work, no machine is ever
going to be able to. Asking them to reason harder or stop reasoning aren't
credible positions.
>Well, you're on solid ground. A resource is identified by URI and may
>emit representations. There's no way to tell from the representations
>what the resource "is"; I tend to believe a resource is what its
>publisher says it is as a good rule of thumb. But it doesn't affect the
>software very much. -Tim
Not web server, DNS, proxies and rendering software, which is the bulk of
the current deployed web as far as I can tell. Any issues or nuances
regarding what the representation in my browser is about or implies are
suitably dealt with my brain. And for the most part I don't need to think
about fine discriminations between things like resources, servers, bits on
the wire and bits on the screen. My brain takes care of it in ways we don't
fully understand.
But it does/will affect any software that is expected to make deductive (or
abductive) inferences about entities using symbols (to avoid the current
terminology debacle). It also affects anyone that is supposed code the
stuff. Sorry, but there is 40 years of applied AI research to testify to
that, never mind a century of mathematical logic. What is required to get
computers to make useful inferences is something we do know a lot about.
The AI/ontology wonks really are not doing this to wind us all up or make us
feel stupid - without the semantics they can't begin to define languages to
let us write this stuff down in a way that is usefully engineered for the
semweb *as posited*. No notation without denotation.
Look, if you or I don't want to deal with the semantic issues because we
believe they're irrelevant to software or even that's it's just impossible
to build a semweb under the state of the art, that's fine, and to certain
degree we'd be correct in either assessement. But to build some kind of
pertinent machinery for /that/ semantic web, we do need formal semantics.
Unfortunately it's more that most of us are prepared to take on; but then
again it's a /lot/ less than the KR folk are confortable with. These guys,
for the most part, are genuinely trying to cut a 70% solution. I'd say most
of us want them to come way under that - their argument is that way under
isn't much use.
Bill de hÓra
|
