RE: Proposed issue: site metadata hook (slight variation)

> > But we are not talking about just another web application. The 
> > Semantic Web is not just a web application. It's a new layer of
> > the core architecture. To compare the Semantic Web with a normal
> > web application is to compare apples and oranges, no actually
> > apples and orange *groves*.
> 
> Excuse me, but that is *NOT* the semantic web vision.
> The semantic web is an extension of the current web to
> include data. 

"extension" != "layer" ???

> This doesn't mean we have to throw away the
> current web.  

I never said we should. You have misunderstood me.

> It is also essentail hat the zexisting web information
> and semantic web information are not compartmentatlized but
> totally integrated.

That is the entire point of MGET and friends. For the interaction
with representations and the interaction with descriptions to
be fully and totally integrated based on common URIs. 

But in many ways, representations are descriptions are distinct
and should be dealt with in a manner that respects that
distinction.

I don't see that as compartmentalization.

> > It certainly is far less reasonable for a given web application
> > to introduce new verbs into the web architecture. I agree. But
> > we are not talking here about just any old web application.
> > 
> > We are talking about the next phase of web architecture, and that
> > warrants new verbs.
> 
> Why?   HTTP didn't merit a new TCP or a new DNS.
> Its strength is that it builds cleanly on top of existing generic
> protools.

This is a pretty weak argument. Just because HTTP didn't merit
a new TCP doesn't mean that the SW doesn't merit extensions to
HTTP (which have no affect on existing web behavior).

HTTP is not intended to be integrated with TCP in the same way
that the SW is expected to be integrated with the Web. 

> > Especially since it is likely, if not certain, that the existing
> > verbs cannot be bent or coerced to work properly.
> 
> You have a jaundiced view of the exsiting verbs, maybe.

I don't think so. At least my view is based on what my understanding
of the specs say.

Yet perhaps therein lies a problem. The specs don't match the
opinion or vision of those who have contributed to them or who 
are tasked to interpret them, and thus there is a tension between
what the community understands and expects, based on the published
specs and what the W3C, TAG, and other "powers that be" intend
and want.

> It is a specific and important goal to build consistently
> with the exusting infrastructure.
> When getting information, use HTTP GET.

When getting *representations*, use HTTP GET. But IMO, that is
too much like a lottery for an effective SW.

If you want a book, HTTP GET it off the shelf. If you want a card 
describing the book, HTTP MGET it from the card catalogue.

> This is basic web architecture.

My point exactly. But it's not sufficient to be the *Semantic* Web
architecture.

> Do do otherwise breaks it.

How does an extension that has no impact on existing behavior
break it?

I guess you feel that WebDAV "breaks" the existing web architecture...

Patrick

Received on Tuesday, 18 February 2003 03:14:04 UTC