Migration from RDSF to OWL was: Re: AS&S and WG consensus

Jeremy Carroll wrote:
> > Yes, well, to speak plainly, I expect that if you we forbidden to use
> > terms "rdfs:seeAlso" and "rdfs:isDefinedBy" under penalty of
> > death, that you
> > would be able to design a long and productive life for yourself. Let's
> > use the term "can't live with" too lightly, eh?
> :)
> you are right of course - (well, God willing).


> but I really do think I would vote against further progress of OWL along
> recommendation track in some instances ... and the RDFS => OWL Lite
> migration is one of those showstopping issues for me (of course that might
> be an incorrect judgement, I am influenced by having worked on RDF
> developers kits and RDF standards; and HP has a greater RDF investment
> say Description Logic investment).

I hope the issue isn't _RDF_ vs _OWL_, rather _RDFS_ vs. _OWL_.

I don't entirely understand _RDFS_ as a standalone 'product'. To me it seems
better presented as a sort of base language onto which one can develop a
production language. With RDFS alone you can't do much, can you? (I am
perfectly willing to be wrong on this issue -- it is merely my impression at
the current point in time). RDFS becomes useful when you add specific terms
(and create something like OWL Full, for example).

In this context I am trying to understand the implications of the migration
path from RDFS to OWL Lite. Can you enlighten me as to how HP is using
_base_ RDFS (i.e. without an 'extension' vocabulary). I mean if an OWL Full
reasoner is so problematic, then wouldn't an RDFS reasoner have the same
exact problem? And if there is no problem developing an RDFS reasoner then
why ought not we use OWL Full, and why not migrate directly from RDFS to OWL
Full? (or whatever parts of OWL Full that one cares to use)


Received on Friday, 24 January 2003 08:46:52 UTC