RDF Conformance was RE: RDF API convergence? was Re: ANNOUNCE: RDF.NET

> Two RDF APIs are considered interoperable whenever:
> 1) System A writes statements in RDF
> 2) System B reads the statements from (1) and stores them in triples
> 3) System B writes the resultant triples from (1) back out in RDF
> 4) System A reads the statements from (3)
> 5) The triples in system A remain the same.

IMNSHO, this is much more important than a standard API. I actually just got
done doing something similar to this at work to prove it was interoperable
with _itself_. We'd serialize our metadata as RDF, read it back into a
separate instance, have that new instance then serialize it's metadata, and
finally compare the two files using diff until they were identical.

What you're suggesting is even more important (and more difficult) because
there will be disparate systems serializing the model. Even if we
canonicalized the XML, the RDF syntax is far too flexible to simply compare
files. Literal objects could appear as either attributes or elements.
Resource objects could be nested inside a property or simply referenced from
it. Do we need a Canonical RDF? If that's even possible. How would you
specify a strict ordering when the RDF Model says order doesn't matter?

What I'd really love to see is a suite of "official" RDF files that can be
used to test not the only the conformance of an RDF processor but also its
"semantic interoperability". It needs to be automatible and it needs to be
the implementor's responsibility to _prove_ that their framework conforms to
the spec by distributing a driver that performs and reports the results of
each of the tests. I believe that all W3C specs should work this way.

I can't help but bring this up but if this had been done for the RDF M&S, we
probably wouldn't have the issues to deal with that we do now. Speaking of
which, Dan, what's the status on the post-Recommendation follow-up you were
going to inquire about for us that will hopefully clear up all of these
issues?

Thanks,
Jason Diamond.

Received on Saturday, 11 November 2000 14:14:05 UTC