Re: new version of rdf:O)-> document

> On May 29, 2009, at 8:36 AM, Peter F.Patel-Schneider wrote:
> 
> > From: Sandro Hawke <sandro@w3.org>
> > Subject: Re: new version of rdf:O)-> document
> > Date: Fri, 29 May 2009 08:34:05 -0500
> >
> >>
> >>>>> (2) The introduction goes on to state that it "does not change the
> >>>>> conceptual model of RDF". This is also not correct.
> >>>>>
> >>>>> At present an API working over RDF which is asked for the  
> >>>>> datatype of a
> >>>>> plain literal should return the programming equivalent of "there  
> >>>>> isn't
> >>>>> one". After the spec such an API should return "rdf:PlainLiteral".
> >>>>
> >>>> Is that true?
> >>>
> >>> Isn't it?  I've lost the plot on what the intention is.  You tell me
> >>> what the working group intends to be the answer here.
> >>
> >> I don't think we have an opinion, since it's an API issue.
> >>
> >>>> My understanding is that it's really up the API and not
> >>>> something that has been standardized.  APIs were always free to do
> >>>> something like this before, and they're free to do something  
> >>>> different
> >>>> even after this (hopefully) reaches Rec.  I guess Jena always  
> >>>> tried to
> >>>> follow the ideas of the spec quite closely, but I don't think all  
> >>>> RDF
> >>>> APIs did, or that the others were wrong for approaching the RDF  
> >>>> data
> >>>> from a different angle.
> >>>
> >>> Sure, that's why I used "should". There is no standardization of  
> >>> APIs so
> >>> each is free to interpret how the formal specs should be  
> >>> manifested to
> >>> the actual users.
> >>>
> >>> That doesn't affect the fact that the conceptual model has changed  
> >>> and
> >>> so APIs are likely to evolve to reflect this. This is hardly the  
> >>> end of
> >>> the world. I just found it hard to accept the bald statement "does  
> >>> not
> >>> change the conceptual model".
> >
> > But the conceptual model has *not* changed.  At all!
> >
> >>> The spec probably does the best that can be done to minimize the  
> >>> impact
> >>> of the change on interoperability.
> >>
> >> How about changing:
> >>
> >>     This extension, however, does not change the conceptual model of
> >>     RDF, and thus does not affect the specifications that depend on  
> >> the
> >>     conceptual model of RDF such as SPARQL
> >>
> >> to:
> >>
> >>     This extension adds an optional element to the conceptual model  
> >> of
> >>     RDF, but does not require any changes to software or affect the
> >>     specifications that depend on the conceptual model of RDF such as
> >>     SPARQL.
> >>
> >> ?
> >>
> >>      - Sandro
> >
> > I am against this change to the document.
> 
> I agree. This kind of micro-tweaking is just going to get people more  
> confused. And strictly, under this proposal, Peter is right: the  
> conceptual model of RDF is not changed.
> 
> What I think is true, and maybe should be said, is that with this  
> particular datatype, the conceptual model **of datatyped RDF** is non- 
> stand... excuse me, unusual, in that this datatype seizes the domain  
> of plain literals for its own syntax space, so that in this datatyped  
> RDF, plain literals are treated as typed. But if you don't use this  
> datatype, RDF is *exactly* the same as it was.
> 
> We could (?) say that  RDF APIs MAY treat plain literals as being  
> identical to typed literals typed with this datatype, in order to  
> facilitate interoperability with tools which actually use the  
> datatype. But its just a MAY.

Those both sound pretty good to me, but I'm easy to please.  Peter,
Dave, are either of them improvements?  Anything else we need to do at
this point?

I'm filing the OWL request for CR today, optimistically including
rdf:PlainLiteral.  My hope is that everyone can sign off on this text by
next Wednesday, and this can go to CR with the rest of the OWL specs.

    -- Sandro

Received on Friday, 29 May 2009 17:52:21 UTC