Re: RDF API

Currently there is no standards-track RDF API. There are however a
number of work-in-progress RDF interfaces, including Sergey's excellent
work which is an outgrowth of his GINF system. Sergey has been working
on a revision of the W3C RDF Parser, SiRPAC, which includes a version
of the proposed RDF API from GINF (replacing a simpler effort than Janne
and I cooked up a year or so ago).

While we should be careful to avoid calling any of the RDF API proposals
"the" RDF API, Sergey's efforts to encourage community input into the
design of that API deserves much encouragement. Just as XML-based systems
that use the community-built SAX interface work just fine, there's every
possibility that the RDF Interest Group might come to an informal
agreement about the shape of a useful RDF API. 

It might be worth giving the thing a name, to avoid any future
confusion. Ron Daniel proposed 'RADIX' a while back (because he thought it
sounded good). I've toyed with 'FAX' by cutesy analogy with SAX, or 'WOM'
for Weblike Object Model by tenuous analogy with DOM). (this would
probably not be a good thread to fill people's mailboxes with though...)

One other thing to say is that the API at 
http://www-db.stanford.edu/~melnik/rdf/api.html is specifically a Java
API. There have been some other RDF systems in (for example) Perl that
expose different styles of interface. Sergey has gone further than anyone
else though in making an explicit proposal to the community of an API for
SAX-like adoption.

All this to say two things really:

 - we should avoid calling anything _the_ RDF API. 

 - we shouldn't let this put us off testdriving such proposals
   and figuring out collectively what works and what doesn't.

So... any comments on Sergey's proposal should go to this list, along with
counter proposals, discussion of APIs in Perl etc. And we'll see where we
can get from here...

Dan

 
On Thu, 23 Mar 2000, McBride, Brian wrote:

> Oops - I may have misinterpreted when I was reading the archive of this
> mailing list.
> 
> What is the status of development of a standard RDF API.  I have been
> working from the one described in
> http://www-db.stanford.edu/~melnik/rdf/api.html.  Is this appropriate?
> 
> Brian McBride
> Hewlett Packard Labs, Bristol.
>

Received on Thursday, 23 March 2000 15:53:33 UTC