Re: How namespace names might be used

At 07:28 PM 6/10/00 -0700, Eric Bohlman wrote:
> > Why can't content negotiation handle this? The CONNEG working group
> > is looking at being able to express just such relationships...
>
>Because content negotiation implies that you connect to a known IP address
>derived from the URI in question and tell it what kind of information you
>want regarding that URI.  IOW, you go to the namespace's owner.  Tim has
>been trying to point out that we need to be able to access useful
>information about a namespace that isn't entirely under the control of a
>single owner.

Interesting assertion.  [Later:  clarified for me by TimBL's comments about 
2-resource and 3-resource uses]

In my work on content negotiation, I have not assumed this is always the 
case (though it is true that some of the work has been informed by that 
kind of scenario).

Certainly there are situations where the CONNEG work can be applied in 
which your assertion does not hold;  IETF work-in-progress illustrating 
this would be the RESCAP protocol, or Presence Protocol.

CONNEG is very consciously just a way of describing combinations of 
capabilities and/or preferences, and does not impose constraints about who 
may ask or who may provide such information, or how.   Similar comments 
apply to the CC/PP work-in-progress.

That said, I think any debate here is about what might be described 
"content negotiation", which is not our central theme.

I was nodding to myself as I read Tim Bray's posting.  In particular, that 
there might be a range of ways to discover semantic information relating to 
a namespace name.  The issue then is to determine under what conditions the 
namespace name is sufficient to somehow indicate the semantic 
information.  I find the approach of David Turner et. al. to be satisfying 
in this respect.

#g

------------
Graham Klyne
(GK@ACM.ORG)

Received on Tuesday, 13 June 2000 08:46:32 UTC