Re: Syntaxes of the mapping language

Simon,

Thanks a lot for this input. I'll take a note and we will certainly use it
in the process of classifying the features (required/time-permitting).
Please allow us some time, as we're currently setting up the infrastructure.

@Harry (our team contact), is the issue tracker already available?

Cheers,
      Michael

-- 
Dr. Michael Hausenblas
LiDRC - Linked Data Research Centre
DERI - Digital Enterprise Research Institute
NUIG - National University of Ireland, Galway
Ireland, Europe
Tel. +353 91 495730
http://linkeddata.deri.ie/
http://sw-app.org/about.html



> From: Simon Reinhardt <simon.reinhardt@koeln.de>
> Date: Thu, 17 Sep 2009 14:06:20 +0200
> To: <public-rdb2rdf-comments@w3.org>
> Subject: Syntaxes of the mapping language
> Resent-From: <public-rdb2rdf-comments@w3.org>
> Resent-Date: Thu, 17 Sep 2009 12:07:03 +0000
> 
> Hello,
> 
> The WG charter contains the following in the scope definition:
> 
> "The mapping language SHOULD have a human-readable syntax as well as XML and
> RDF representations of the syntax for purposes of discovery and machine
> generation."
> 
> In my opinion this is way too many representations (given that RDF itself has
> dozens of syntaxes already) which will be a huge burden on the implementors.
> 
> I think OWL 2 has way too many representations already just so everyone gets
> their own preferred way of doing things - at the cost of OWL implementations
> which have to support as many syntaxes as possible in order to be able to
> parse ontologies found on the Web. For SPARQL it is similar, there are many
> representations of the query results. The reason here though is to make
> implementations of *consumers* easier because you can use technology available
> in the browsers for creating mash-ups and don't have to care about the
> Semantic Web background too much.
> 
> The question is then: how and where will the mapping language be used? Is it
> for server-side tools only which have access to the database and do the
> transformation to RDF? Or is there any use-case where you want to publish the
> mappings on the Web and client applications can retrieve them and do something
> cool with them?
> 
> I'd settle for one syntax/representation only. If you make it RDF you get nice
> interoperability with the rest of the Semantic Web toolchain but is that
> really needed? Depends on the use-cases. And working with RDF is not always
> easy because it's so unrestricted. Having a human-readable syntax is a good
> goal but that doesn't exclude XML. If it's a well-designed syntax (bad
> example: XSLT ;-) ) people can work with it easily enough - they're used to
> XML for all sorts of configuration files and you have excellent tool support
> for it.
> 
> And then of course you have to look at the existing stuff. But I'd rather you
> come up with something new than introduce too much variety.
> 
> Thanks,
>   Simon
>

Received on Friday, 18 September 2009 08:12:52 UTC