RE: ISSUE-137 (including XML includes)

Hello,

>From the current discussion, it seems to me that there is a strong need for RDF includes -- that is, the ability to include (i.e.,
embed) an RDF graph into another graph. This feature should also support various RDF syntaxes.

I find this potentially quite useful; however, I don't think that OWL is then the correct forum for designing such a feature: this
should be done by an RDF Working Group. We can't rewind the clock and add this feature to RDF now; however, we should not now
completely contort the various layers in the Semantic Web. In particular:

- XML includes are out there, so one can use them for including RDF/XML graphs into RDF/XML graphs. This should be sufficient for
quite a few applications, as RDF/XML *by far* the most common exchange syntax for RDF-based data nowadays.

- We should provide a feature request for a future revision of RDF in form of syntax-independent graph inclusions.

This seems to me to be a clean solution in which different concerns are addressed at appropriate layers. 

Regards,

	Boris

> -----Original Message-----
> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org] On Behalf Of Sandro Hawke
> Sent: 12 September 2008 16:27
> To: Bijan Parsia
> Cc: Alan Ruttenberg; Peter F. Patel-Schneider; public-owl-wg@w3.org
> Subject: Re: ISSUE-137 (including XML includes)
> 
> 
> >
> > On 12 Sep 2008, at 16:05, Alan Ruttenberg wrote:
> >
> > > On Fri, Sep 12, 2008 at 6:48 AM, Peter F. Patel-Schneider
> > > <pfps@research.bell-labs.com> wrote:
> > >
> > >> Isn't RDF/XML *the* RDF serialization?
> > >
> > > That we have turtle in the Primer is just one piece of evidence that
> > > this is not the case. As I said, and as had been argued persuasively
> > > in the past, a solution that is specific to a particular serialization
> > > of RDF is undesirable.
> >
> > We don't spec a Turtle serialization of OWL.
> >
> > It seems very odd to constrain ourselves from using standard, widely
> > implemented, W3C technologies as they are intended to be used in
> > order to accommodate a non-normative syntax over which we do not
> > have, nor do we desire, any control.
> 
> I don't know what the issue here it, but on this point I'll jump in:
> it's been clear since at least the 2001 charter for RDF Core that
> RDF/XML was not intended to be the only standard serialization of RDF.
> RIF BLD frames are probably a Rec Track serialization of RDF, and Turtle
> is certainly a de facto standard serialization or RDF.
> 
>     -- Sandro
> 

Received on Friday, 12 September 2008 15:38:39 UTC