Re: Are MGET descriptions workable/necessary?

On Monday, Nov 24, 2003, at 00:02 Europe/Helsinki, ext Phil Dawes wrote:

> Hi Patrick,
> Patrick Stickler writes:
>> [...]
>> It's no more brittle than the web is.
>> If you have a URI and you want a
>> representation of the resource denoted by that URI, you ask an HTTP
>> server hosted at (which is presumed to exist) and
>> usually, you'd GET back a representation.
>> If you want a description of the resource denoted by that URI, you
>> ask the HTTP server hosted at, and if that server is
>> URIQA enlightened, you'd MGET back a description.
>> [...]
>> If MGET is brittle. Then so is GET.
> I agree that the mechanisms are the same. It's actually the social
> burden on the term author that I'm not convinced about. The difference
> is that a web page, being a non-authoritative representation, can be
> moved around, 302'd, re-directed via an html link, update your
> bookmarks please'd, and eventually retired.
> An RDF term is forever. Just think about that.

I'm sorry, but I don't see any difference. Cool URIs don't change, 
whether one uses them to interact with representations or descriptions.

If a resource "moves", it gets a new URI, and the same redirection that
occurs for GET requests would have to happen for MGET requests.

> In my lifetime I'll be
> creating probably millions of terms (already been responsible for
> thousands in my work intranet). In order for this authoritative
> description mechanism to work, I'll need to maintain an http service
> for each of the terms I create *at the URL I mint it at!!!*

I think you are artificially exaggerating the support for descriptions
over the support for representations. I see them as being equivalent.

> And so will *everyone* else!
> I can't split off a term and give it to somebody else to maintain,
> because it's tied to my domain name.
> That means I've got to deal with load, infrastructure, dns expiry
> etc.. etc.. forever!

Er. Yes. That is the nature of URIs.

*However* if you are concerned with long term "portability" of
URI ownership, you can employ various means to create URIs with
minimal mnemmonic content, which are hosted in a neutral context,
i.e. PURLs (of various sorts).

These issues are IMO completely orthogonal to the discussion at hand.

> That makes it wholly unreliable in my opinion, and not the sort of
> thing to be bootstrapping the SW with.

Well, then I wonder whether you fully understand what it means
to bootstrap the SW.

If one *does* have a URI via which one can get representations, how
else does one get a description of that resource without any other
knowledge *except* via that URI?!

You seem to be seeing monsters in the shadows when all they
are are shadows...


> Best regards,
> Phil

Received on Monday, 24 November 2003 12:03:20 UTC