Re: RDFa Core and RDF/XML

Hi Ivan,

> SVG does not talk separately about an RDFa processor. The complete, relevant quote is:
>
> [[[
> SVG provides two mechanisms for adding metadata directly to a document: the 'metadata' element, and several metadata attributes. These different mechanisms may be used independently, or in concert. If both are being used for RDF (that is, RDF and RDFa), then an RDF processor should combine them into the same graph. (Note: metadata attributes should not be used directly on RDF elements.)
> ]]]
>
> which, I think, has the intent of what I described in terms of an RDFa processor. But I am not an SVG WG member:-)

Right. But an 'RDF processor' is very different to an 'RDFa processor'.

Since RDF must be manifest in some serialisation, then technically
there is no such thing as an 'RDF parser' -- only RDF/XML parsers, n3
parsers, RDFa parsers, and so on. So although there is no definition,
I'd see an 'RDF processor' as something that supports one or more
parsers.

But an 'RDFa processor' is really just an 'RDFa parser' -- it's
difficult to see what else it can mean.


> My intention was indeed that an *RDFa* processor should parse the RDF/XML portion, too,
> and add it to the RDF graph that is produced. Writing these lines made me realize, however,
> that this *is* an extra load on an RDFa processor, so it is not so obvious after all:-(

I think it's conceptually wrong, too. Whilst an abstract 'RDF parser'
might process many serialisations, it would be wrong to expect a
parser that handles one serialisation to take on board other
serialisations.

(Which in passing is one of the reasons for introducing the notion of
the 'default graph' into RDFa; it gives us a fighting chance of
ensuring that all triples from all parsers end up in the same place.)


> It is certainly true that an engine extracting metadata from an SVG file that wants to be
> conform to the SVG standard has to do both the RDF/XML and the RDFa processing. But
> that may be then a tool's own option (my distiller does that).

Right, but that would be an 'RDF processor' that just happens to
incorporate both an RDF/XML parser and an RDFa parser.


> Whatever we decide, we may want to add an informal note to the RDFa 1.1 Core about this.
> It is, I think, an obvious questions user may ask themselves.

If we did make such a note, perhaps it would be in relation to notions
of the default graph, rather than processors? That's the main way that
different parsers can align their triples.

Or maybe we clarify this point about an RDF processor containing one
or more parsers, and say that all parsers running in the context of
the same RDF processor should ideally populate the same default graph.

Regards,

Mark

--
Mark Birbeck, webBackplane

mark.birbeck@webBackplane.com

http://webBackplane.com/mark-birbeck

webBackplane is a trading name of Backplane Ltd. (company number
05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
London, EC2A 4RR)

Received on Monday, 29 March 2010 13:03:44 UTC