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

> -----Original Message-----
> From: ext Jeffrey Winter []
> Sent: 12 February, 2003 17:02
> To: Stickler Patrick (NMP/Tampere);; 
> Subject: RE: Proposed issue: site metadata hook (slight variation)
> > But, again, this imposes two system calls to get descriptions
> > of resources *and* requires the explicit naming of those
> > bodies of knowledge, which is unnecessary.
> I don't see the two system calls as being particularly 
> problematic since
> in the general case, "metadata" resources would be requested 
> orders of 
> magnitude less often than "normal" resources. If use cases could be 
> described where it was an issue I'd be interested in seeing them.

I think that once rich metadata is available in volume, once a
reasonable architecture is in place to serve it, that *most* 
automated agents and web services will be wanting metadata alot.

> I also don't see why you *wouldn't* want to make the metadata, in any 
> form, an addressable resource.  

I'm not saying you shouldn't be able to, only that in most use
cases I can dream up or have heard, you wouldn't want or need
to. When's the last time you've referred to "all the knowledge
known by X about Y"?

And if you want to, fine, give it a name and do so. But don't
*require* the duplication of URIs on the web just to allow
one to get descriptions of resources.

> I take issue with the original premise the that metadata should 
> ever be non-addressible.

I've never presumed nor asserted that. Ever. Go re-read my posts.

I've only ever said that one should not need a *different* URI
from that of the resource in order to publish or obtain or
otherwise interact with descriptions of that resource.


Received on Wednesday, 12 February 2003 10:19:31 UTC