[Mulgara-dev] descriptors/references

Paul Gearon gearon at ieee.org
Mon Jul 16 16:22:56 UTC 2007


I'd also like to point out that RDF uses URIs and not URLs.  This
means that it contains identifiers and not locations.  The fact that a
URI can be a location is both useful and unfortunate.  :-)

Namespaces are specifically not required to refer to a document.
However, I've always been pleased that they DO, as the information
found there is usually quite useful.  In line with the discussions
that David mentions, I've noticed that many people are expecting to
find documents at the locations described by namespaces that happen to
be URLs, and there is a move to formalize this.

Paul

On 7/16/07, David Wood <dwood at softwarememetics.com> wrote:
> Hi all,
>
> To be clear, I am suggesting that *neither* the RDF nor Mulgara
> namespaces should return a document.  Instead, both namespaces should
> return an HTTP 303 status code that may optionally point to documents
> describing the resources.
>
> Regards,
> Dave
>
>
> On 16 Jul2007, at 10:02, David Wood wrote:
>
> > Hi Bill,
> >
> > On 7 Jul2007, at 01:05, William Mills wrote:
> >> So the major difference between rdf and mulgara that I have found
> >> so far is
> >>
> >> That http://www.w3.org/1999/02/22-rdf-syntax-ns is actually
> >> available as an XML reference document, and http://mulgara.org/
> >> mulgara is not.  Should this be true?
> >
> > It depends and is under active discussion, especially within the
> > W3C Technical Architecture Group.  At issue is whether HTTP URIs
> > may refer to things that are not information resources (such as
> > physical resources or namespaces) and whether the term "information
> > resource" includes virtual concepts like namespaces at all.  This
> > discussion is complicated by the fact that "information resources"
> > are currently defined as things you get back from a URI
> > dereferencing when the server responds with a 2XX HTTP response
> > code (!).  It is all very circular.
> >
> > The current state of thinking is that HTTP should return different
> > status codes depending on what you are trying to do.  This should
> > eventually result in a clean handling of both information and non-
> > information resources by HTTP (and hopefully encourage everyone,
> > including Mulgara) to ground their namespaces in URIs.
> >
> > In brief, an HTTP 2XX return code should result when the
> > transaction returns an information resource.  Non-information
> > resources should result in an HTTP 303 (See Other) response, which
> > may optionally list one or more URIs that relate to the non-
> > informational resource (but are not in fact the resource itself).
> >
> > Brian Sletten of the Mulgara community and I are actively upgrading
> > the Persistent URL or PURL service at http://purl.org to handle the
> > full range of HTTP return codes for this purpose.  It is our hope
> > that Mulgara will follow others in using PURLs for namespace
> > dereferencing.
> >
> > Here is the current W3C guidance, but note the large number of
> > disagreements in the draft document:
> >
> > Dereferencing HTTP URIs
> > Draft TAG Finding 31 May 2007
> > http://www.w3.org/2001/tag/doc/httpRange-14/2007-05-31/HttpRange-14
> >
> > and the older open TAG issue that started this discussion:
> > http://www.w3.org/2001/tag/issues.html#httpRange-14
> >
> > Regards,
> > Dave
> >
> >
> >
>
> _______________________________________________
> Mulgara-dev mailing list
> Mulgara-dev at mulgara.org
> http://mulgara.org/mailman/listinfo/mulgara-dev
>



More information about the Mulgara-dev mailing list