RE: Proposed response for timbl-02 (reification semantics)

> -----Original Message-----
> From: ext Brian McBride []
> Sent: 09 April, 2003 15:41
> To: Graham Klyne;
> Subject: Re: Proposed response for timbl-02 (reification semantics)
> Graham,
> This looks fine, with a couple of caveats.
> I note from Tim
> [[
> As Director, I wonder about whether the group can claim this part of
> the spec to have reached its implementation requirement,
> if the parsers parse the information but the semantics have not been
> field tested.
> ]]
> We had three folks say at the telecon that they used reification as 
> defined.  Mike has written his up.  I recall that Patrick and 
> Frank were 
> the others.  We also have the p3p rdf schema.  I suggest we also ask 
> Patrick and Frank to say document their use cases on the 
> record so we can 
> point to them from the WG response.
> Patrick, Frank - you ok with that?

I'm fine with that, but our use is internal, so I can't provide links.

But in a nutshell, we use reification to record the source of statements,
by which we infer authority. E.g.

   #subj #pred #obj .

   _:s rdf:type rdf:Statement ;
   _:s rdf:subject #subj ;
   _:s rdf:predicate #pred ;
   _:s rdf:object #obj ;
   _:s #source <http://some.server/some_schema.rdf> ]

   <http://some.server/some_schema.rdf> dc:owner #auth .

We can then qualify queries based on either source or authority.

Shortly, when I get online and functional (a week or two) this will
be evident in the public metadata published there.

Hopefully the above summary suffices for the record, re the WG's decision
on this issue.

> That would still leaves Tim's point
> [[
> (Remember the story of the man who wrote make(1) and a few
> days later realized that the tab/space distinction in the Makefile
> syntax was a mess, but didn't like to change it because by that time
> several of his colleagues were using the syntax?)
> ]]
> which I read as saying that for the long term good of the 
> semantic web we 
> should ignore the short term pain.  Tim will have the option 
> of coming back 
> and saying that, but we will at least have differentiated 
> between there 
> being no use (which is what I think Tim might believe) and 
> not enough use.

Well, if the WG chooses to remove the reification vocabulary, we will
simply have to define it ourselves, which is not hard to do, but
then, there at least is a degree of reasonable interoperability
afforded by the present vocabulary, and I also think a good bit
of concensus about its usage.

I would hate to see it go (but could live with it going).

> Also, I think we agreed to put health warnings on semantics 
> and schema so 
> that folks were aware of what reification isn't appropriate for.

I fully agree. We should be clear about the various uses we identified
in our discussions and clarify which use we chose as the intended use
and warn folks not to use it for the other uses.


Received on Thursday, 10 April 2003 02:50:12 UTC