Phil Archer | 2 Nov 23:07 2007

Re: Pointer to the ontology for generic resources. HTTP Link header.

There's a good deal of 'noise' about the HTTP Link header at the moment 
- and not all of it created by me (honest!).

In the currently published draft of the POWDER Description Resources 
document [1] we posit the use of HTTP Link - but, as the doc shows, this 
is something of an open question. Dan Connolly [2] questioned it and, 
well, there's a short thread with lots of links to evidence to support 
the contention that the only real problem with the HTTP Link header is 
that, whilst it was defined in RFC 2068, it was dropped in RFC2616 (but 
it remains in the registry).

The new HTTPBIS WG [3] isn't going to re-instate the header, but Mark 
Nottingham may re-issue his draft on this and the associated HTTP 
Profile header [4]. The 'protocol part' of the Protocol for Web 
Description Resources is handicapped by its absence so I'm certainly 
hoping Mark will re-issue it and that there will be support from 
elsewhere. I guess it's all about trying to build momentum, show use 
cases and the fact that, well, it's proven to work.

On a practical level, the dependency on HTTP Link being formalised has 
been removed from the next published draft of the POWDER doc but we hint 
at its existence and plan to show example of its use in the (yet to be 
written) non-normative Primer.


(Continue reading)