Re: HTTP in RDF

>
> Another thought about HTTP in RDF
>
> <http://lists.w3.org/Archives/Public/public-wai-ert/2005Nov/0022> and
> <http://lists.w3.org/Archives/Public/public-wai-ert/2006Jan/0035.html>:
>
> Originally, I created different namespaces for HTTP and each of the
> origins for extension headers (rfc822, rfc2295, rfc2965 and nsCookie).
> This would be more difficult to implement, because you have to look for
> the schema a specific header in the request or response is defined in
> before you create the RDF properties / XML elements. It would be much
> easier to have them all in one namespace. Do we gain anything important
> when we have different namespaces?
>
> Another thing: when there is only one namespace, the problem with the
> two Cookie header definitions is gone ...

You seem to have talked about this topic during the last telcon.
Unfortunately, I don't understand the "two-level namespace" stuff.

So let me explain my proposal.

HTTP does not use any namespaces. For HTTP an extension header is just a
header like a standard one. So when you see a header, you don"t know, if
it is a standard header or an extension. Of course you can maintain
several lists of headers defined in different sources to then map the
header to some "namespace". But then you would introduce a concept that
does not exist in the original.

I propose to use one namespace for all the named header properties. A
header for which there is currently no property listed in the NOTE, would
be modelled with the rfc822 scheme. But the list of defined header
properties could be extended over time so that more extension headers
could be recognized directly.

Is that a feasible approach?
--
Johannes Koch - Competence Center BIKA
Fraunhofer Institute for Applied Information Technology (FIT.LIFE)
Schloss Birlinghoven, D-53757 Sankt Augustin, Germany
Phone: +49-2241-142628

Received on Tuesday, 7 February 2006 21:54:13 UTC