[foaf-dev] [FOAF spec revised] foaf:logo an inverse-functional-property?

Hogan, Aidan aidan.hogan at deri.org
Wed Aug 11 01:22:12 CEST 2010


> I've just committed a revision to the FOAF spec. It's dated 9th August
> and minor tweaks are (as usual) quite likely for the next ~24 hours.

Hi Dan,

CC'ing the pedantic-web mailing list [might be some useful opinions on
there about the first issue].

>     * index.rdf: foaf:logo is an Inverse Functional Property now"

This worries me somewhat in that foaf:logo is already in widespread use
and I can think of examples where this might cause problems. For example
consider the logo of "The Coca-Cola Company" and "Coca-Cola" the
product: they're feasibly the same, no? For me, even if I knew foaf:logo
was an inverse-functional-property (and only assuming I had one
Coca-Cola logo image), I'd have a hard time deciding which deserved it
more: the company or the product.

For me, inverse-functional-property declarations should be completely
watertight: otherwise, the Web will find ways to create
counter-examples. From [1], I managed to find one or two groups of
resources which share logos and aren't really the same [2,3] -- one
could argue that these usages of (the previously vague) logo property
were valid at the time. (That said, I also managed to find some
resources that share logos that are equivalent.)

Any thoughts on this? If you do leave foaf:logo as
owl:InverseFunctionalProperty, you should look at updating [4] with some
discussion.

I have one other *small* issue might be worth looking at for the next
version of FOAF, relating to rdfs:label values for terms. This is
perhaps being way too pedantic, but any chance of making the labels more
consistent in terms of presentation. For example, foaf:surname currently
has uppercase "Surname", and foaf:givenName has "Given name", whereas
foaf:lastName has "lastName". This makes domain-agnostic rendering of
FOAF a little prettier [5]... not by any means a priority of course.

Also couldn't resist asking:

>     * index.rdf: removed some false disjointness claims: people can be
> documents (tatoos);

...anything to do with [6]? Or perhaps you've just been watching re-runs
of Prison Break :).

Cheers,
Aidan

[1]
http://lod.openlinksw.com/sparql?default-graph-uri=&should-sponge=&query
=SELECT+%3Fs1+%3Fs2+%3Fo+%3Fc1+%3Fc2+WHERE+{GRAPH+%3Fc1+{%3Fs1+foaf:logo
+%3Fo+.}+GRAPH+%3Fc2{+%3Fs2+foaf:logo+%3Fo}+FILTER(%3Fs1+!%3D+%3Fs2)}&fo
rmat=text/html&debug=on&timeout=

[2]
http://creativecommons.org/licenses/by-nc-sa/3.0/ro/rdf
http://creativecommons.org/licenses/by-nc-sa/3.0/rdf

[3] 
http://data.openlinksw.com/oplweb/product_family/development#this
http://data.openlinksw.com/oplweb/product_family/professionalservices#th
is
http://data.openlinksw.com/oplweb/product_family/uda#this
http://data.openlinksw.com/oplweb/product_family/utilities#this
http://data.openlinksw.com/oplweb/product_family/virtuoso#this

[4] http://xmlns.com/foaf/spec/#term_logo

[5]
http://swse.deri.org/detail?focus=http://sw.deri.org/~aidanh/foaf/foaf.r
df%23Aidan_Hogan 

[6] http://www.slideshare.net/aidhog/weaving-the-pedantic-web-ld (slide
21)



More information about the foaf-dev mailing list