Re: [foaf-dev] Implying that two mentioned individuals are distinct... updating best practice from owl:differentFrom to owl:IrreflexiveObjectProperty?

Dan,

<spiel>
We work on reasoning and entity consolidation (smushing) over arbitrary
RDF Web data, and I personally encourage any reasonable amendments to
FOAF which will help us to automatically identify noisy inferences in
the data. This particularly means anything that will flag any obvious
noise as inconsistency, which we can automatically detect and repair,
and which ultimately improves the quality of reasoning results.

Currently, there aren't enough indicators of when something goes wrong
in the popular vocabularies, and the current level of granularity for
diagnosing problems is pretty crap. If you've no way of cross-checking
your results, then automatic reasoning and consolidation becomes a lot
more hit-and-hope.

I'm not talking about adding endless constraints about how the
vocabulary can be used, but talking about simple axiomatisations of
things that are obviously noise. For me, the recent removal of
disjointness contraints from FOAF is a step in the other direction,
though I understand where you're coming from. In my opinion, something
being simultaneously a Person and a Document is an "obvious" indicator
of noise, and really does help in automatically diagnosing problems.
</spiel>

> He suggests http://www.w3.org/TR/2009/REC-owl2-syntax-
> 20091027/#Irreflexive_Object_Properties
> will do the job, and that this doesn't put the idiom into OWL Full.
> 
> So we would on this usage assert that rel:livesWith, foaf:knows and
> similar relations are owl:IrreflexiveObjectProperty ie. that they
> can't truthfully be used to assert of some resource r that it stands
> in that relationship to itself.

Although foaf:knows could be interpreted as possibly being reflexive, I
haven't seen such usage in the wild, and would definitely encourage this
addition---however, you should make the irreflexivity clear in the
human-readable spec.

But why stop there? Why not give reasoners more sway in identifying
obvious noise? Look at every property of FOAF and mark all obvious
asymmetric/irreflexive properties as such. (Note that it may be
redundant for certain properties whose domain and range are already
disjoint, but not for other properties---a little redundancy is not the
end of the world in any case.) 

> Ian - do you have the impression that anyone is using the
> owl:differentFrom claims, or they serve essentially as (language
> neutral) documentation of your intended meaning?

A minority of FOAF-hackers are using/have used owl:differentFrom to
counteract incorrect consolidation they've identified for themselves.
These include Richard (not anymore?), Sergio [1], and yourself of course
(though you should be more concerned about Domenico Gendarmi than about
your namesakes---Domenico has ostensibly been sharing your w3.org email
with you [2] o.O).

Again, anything you can add to FOAF which helps to identify clear-cut
noise is appreciated.

Embrace inconsistency---it is much needed detergent for the Web.

Cheers,
Aidan

[1] http://www.wikier.org/foaf.rdf
[2]
http://foafbuilder.qdos.com/people/myriamleggieri.wordpress.com/foaf.rdf
(362ce75324396f0aa2d3e5f1246f40bf3bb44401)