Re: RDF/XML Syntax Specification (Revised) W3C Working Draft published

On Tuesday, December 18, 2001, at 08:06 , Dave Beckett wrote:

>
> I'm pleased to announce that a new version of the ongoing work on the
> RDF/XML syntax has been published by the W3C RDF Core Working Group,
> as part of the Semantic Web Activity.
>
>   RDF/XML Syntax Specification (Revised)
>   W3C Working Draft 18 December 2001
>   http://www.w3.org/TR/2001/WD-rdf-syntax-grammar-20011218
>
>
> The most significant changes since the 06 September 2001 working
> draft are:
>
[ ... ]
>   * The grammar was updated to remove rdf:aboutEach
[ ... ]
> The working group in particular want to point out the removal of
> rdf:aboutEach from the language and solicit feedback on this.  In
> summary it was deleted because:
>   - it is not used
>   - it is not widely implemented
>   - it has confusing interactions with rdf:bagID
>   - it does not scale as parsers have to save state
>   - this is the wrong layer in which to implemenent such functionality
>
> The background to this will be recorded at
>   http://www.w3.org/2000/03/rdf-tracking/#rdfms-abouteach
> in due course, but you might also want to look at
>   http://www.w3.org/TR/2001/WD-rdf-syntax-
> grammar-20011218/#rdfms-abouteach
> for pointers.
>

My main concern with this is what to use instead. As I mentioned 
previously, I'm currently defining an internal standard, or template, 
for inline metadata. This metadata will be partially manually generated, 
and so shouldn't be unnecessarily verbose. A little syntactic sugar 
would be handy.

I'm not tied to aboutEach, since there is no prior use. I am looking for 
best practice here, so as to start off in the best way. I could quite 
easily do a little XSLT preprocessing to convert uses of aboutEach to 
another form, but it
seems silly to do this, if an alternative exists.

To widen this a little: how is is RDF to be used by people who have 
domain knowledge, but little RDF experience, and do not wish to write 
screeds of repeated verbose data?

I am quite keen on using RDF for `proper' knowledge representation, but, 
to be honest, the average user runs to the hills when they see a 
slightly non-trivial usage of RDF (especially as RDF/XML). Is RDF even 
*meant* to be manually generated? If so, it seems natural to those of us 
coming from a programming background that it should support minimisation 
of input, where shared properties exist eg if Tom has properties A, B 
and C, while Jane has properties C, F, and A, it is automatic to group 
the shared properties together and mention them first, then go onto the 
differences. This syntax does not have to be reflected in the model, it 
is merely a short-hand, but a very valuable short-hand.

I'm an RDF newbie, so I may be missing some things, but it seems this is 
an issue if RDF is be a human-writable language.

--
Mike

Received on Wednesday, 19 December 2001 05:02:56 UTC