Re: Publications on SOA and linked data?

On Mon, Aug 24, 2009 at 2:31 PM, Axel Rauschmayer<axel@rauschma.de> wrote:
> Web services and linked data seem highly related: Many of the linked data
> introductions feel ReSTful, as does Tabulator's use of SPARQL/update. But,
> while there are many blog posts out there that briefly touch on this topic,
> I have yet to find a publication that paints a complete and coherent
> picture. Is anyone aware of such publications (or currently writing one ;-)
> ?
>
> There are semantic web services, but I would expect linked data web services
> to be different.

There were some discussions a long while back (~2001, while SOAP was
being W3C-standardised), on two themes:

- SOAP's "SOAP Encoding" data model (for serialization of
edge-labelled data graphs) and RDF
- Use of SOAP for REST services (versus ignoring it)

I made some experiments back then with REST-style publication of RDF
graphs that were serialized using the SOAP Encoding XML syntax instead
of RDF's, and Max Froumentin wrote an XSLT that turned SOAP Encoded
graphs into RDF/XML.

There is a (rather drafty but perhaps helpeful) document on all this
at http://www.w3.org/2001/sw/Europe/reports/xml_graph_serialization_report/
- that is quite wordy, but might help explain some of the differences
in approach and terminology between the web service and SW
communities.

For more discussion eg. see
http://aspn.activestate.com/ASPN/Mail/Message/xml-dist-app/1347476
http://aspn.activestate.com/ASPN/Mail/Message/xml-dist-app/805271

For those of us who care about data linking, REST-style publication
has a lot of appeal and REST is our general working assumption on how
to use the Web. A great many ws-* Web services have offerings can be
seen as simple lookups (eg. temperature given a city code) which map
into URL patterns. So it might be that many of the "Linked Data" Web
services you're looking for, are out there, but are cleverly disguised
as ordinary old-fashioned Web sites, only with nice clean URI
structures, content-negotiable formats, alongside a few REST lookup
services or a SPARQL db. A lot of the great work coming out of the BBC
lately falls under this heading imho - the Web "service" simply is the
(nicely designed) Web site. Lookups correspond to pages, and various
content formats (HTML, RDFa, RDF/XML, JSON) carry the responses to the
"api calls".

cheers,

Dan

Received on Monday, 24 August 2009 13:09:09 UTC