W3C home > Mailing lists > Public > public-rdf-wg@w3.org > February 2012

Re: Another try.

From: Eric Prud'hommeaux <eric@w3.org>
Date: Tue, 21 Feb 2012 10:17:09 -0500
To: Andy Seaborne <andy.seaborne@epimorphics.com>
Cc: public-rdf-wg@w3.org
Message-ID: <20120221151706.GT28311@w3.org>
* Andy Seaborne <andy.seaborne@epimorphics.com> [2012-02-21 11:44+0000]
> 
> 
> On 21/02/12 07:56, Pat Hayes wrote:
> >For conservatives among us, the opposite re-interpretation is always
> >available. Any quad-graph can be thought of as a SPARQL dataset, by
> >'slicing' the quads according to their last argument, and
> >re-declaring this parameter to be a graph label. However, to retain
> >the semantic flexibility (ie to have the triples in each graph able
> >to be re-interpreted differently in each labeled graph), we would
> >have to modify the RDF semantics to allow for this graph-local
> >context being involved in the truth recursions. And as already noted,
> >it is simpler, and much less of a change ot the basic RDF model,  to
> >do this by thinking of this construction in the quad-graph way as
> >being a set of property-with-three-argument quads rather than as a
> >collection of labelled sets of two-argument triples. And as so many
> >of the 'natural' uses of datasets seem to want to take advantage of
> >the apparent contextual' possibility of the graph label, and this
> >option is only available in a quad-store format in any case, it seems
> >comparatively harmless to attach the needed semantics directly to
> >this quad store format, rather than tinker with the semantics of
> >triples or try to make sense of graph 'names' which do not denote
> >graphs.
> 
> I was wondering about existing vocabularies.
> 
> If I understand the quad proposal, then all existing vocabularies
> are technically undefined because they never define P(S,O,G), only
> P(S,O).

And, imagining nimble fingers extended these definitions to include
P(S,O,G), could they construct a definition for P(S,O,_) where the
predicate means the same thing *in* *every* *graph*? This is of
manifest importance; why else would we even write the definition.
Enabling this, however, flies again in the face of those who want to
"contextualize" the interpretation of graphs. The ability to say
foaf:mbox is "an Internet mailbox associated with exactly one owner"
is fundamentally in tension with wanting to say that foaf:mbox is
context-dependent, and only the former leads to global knowledge
exchange.

If I understand the goal of the contextualist argument, is that the
context influences our belief in assertions. Named graphs are a way of
coursely controlling the assertions we let into our universe, and even
predicating confidence in them. E.g. I'm trying to send PatH some mail
and I'll presume the accuracy of the foaf:mbox in a FOAF file that I
believe he controls and maintains. It turns out PatH uses mboxMD5sum
instead (allergy to spam) so I look in the FOAF docs of people that I
know because I don't trust other folks not to lie.

In biology, as I expect in any detailed discipline, confidence is
considerably more nuanced than that. I could have a typical database:
{
    <http://w3.org/brief/MjU2> :mismatchesInCitation 0 .
}
<http://w3.org/brief/MjU2> {
    human:OR4F5 :homologousTo mouse:Olfr1288 .
}
and ask for human genes which are exactly like mouse genes (0
mismatches). This example is as simple as possible, but still has
domain details. I suspect that 80% of our potential use cases are
similarly detailed in some domain.

Almost every question I ask in anger is one which is predicated on
some context, but that doesn't mean that I want the interpretations of
a predicate, e.g. :homologousTo, to vary with context. The semantics
must enable predicated trust, but I suspect need not wrestle with the
fact that there may be mutually incompatible assertions in the world.
In my ignorance, I read PatH's proposal as trying to deal with
potential incompatibilities by undermining the strength of their
assertions rather than leaving that up to the consumer of the data.


> The graph-local context view seems to preserve the vocabulary by
> reinterpreting P(S,O).  It's less neat to have sets of triples +
> graph labels, but it does seem to carry-over existing data.
> 
> Have I missed something?
> 
> 	Andy
> 

-- 
-ericP
Received on Tuesday, 21 February 2012 15:17:42 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:47 GMT