Dark triples: call for order

Gentlepeople, Guus already sent out an important remark [1] on the
need for clarification about the dark triple issue.
But from the first replies, there's the risk we just go on with
the discussion that has taken the wg in the swamp.
Taking a look at the archive....
** 39 emails in the last two days **
(could have missed some, blub...)
39 emails in two days are too much, if we don't have a clear direction
of where we're going, and what our primary goals are.

Our *primary goal* is to clarify the supposed dependency that RDF-Core
should have on the dark triples issue. We have asked RDF-Core for an
important modification of the current RDF model, and we've been
(correctly) pushed back, saying that before issuing such a critical
request, we'd give precise motivations and very good reasons to do so.

Therefore, what the dark triple task force (with help of the rest of the wg)
should do is not to try to give a solution to entailment problems, paradoxes
and so on. Because, yes, these might be (are!) a problem, but just for OWL.
The first primary goal of the task force should be just to determine
whether there is some problem with RDF, or it is instead the case that
these are just our (WOWG) problems: maybe hard to solve (how much a
powerful language we want, with cycles or not, with paradoxes or not),
but in any case problems that do not impact RDFCore.

There's no easy way out to the "language expressiveness" problem, that's
something that needs discussion (and maybe lot of emails...). But that's
something
we can do later, after having clearly separated the expressiveness problems
on one side, and the RDF-model problems on the other.

This will enable to solve the critical issue of whether we've to ask RDFCore
to actually take the hot potato (and so, having their timeline critically
affected) or not.

So, summing up, let's focus on the "why in earth RDFCore should do some
modification to the RDF model?" problem. Clear examples, clear motivations
on why the problem can't be just solved in OWL (with zillions decisions and
problems, but just in OWL). And if we want, but just optionally, even clear
definitions
of possible solutions for RDFCore to consider (like dark triples).

RDFCore is waiting for us, and unless we show them that RDF has
*necessarily* to be
changed, the hot potato is ours, not theirs ;)

-M

[1] http://lists.w3.org/Archives/Public/www-webont-wg/2002Apr/0233.html

Received on Tuesday, 23 April 2002 12:20:56 UTC