[rdfweb-dev] weblogs and homepages

Dan Brickley danbri at w3.org
Tue Jun 24 22:53:07 UTC 2003

thanks for the writeup! Just a brief reply for now. I think the key to 
getting this right is to note that only some Web logs are the foaf:weblog of
a Person; and that a person can contribute to a weblog, yet it not be 
'their' foaf:weblog. Same with homepages. Most widespread are those of 
people and companies and projects; each is the homepage of something. Also 
we  see homepages of couples, of pets, etc. In all these cases, we should 
defer to their self-descriptions, ideally. If a page says that it is the 
foaf:homepage of a thing that is a wordnet:Couple, or a wordnet:Cat, then so
be it. All the owl:InverseFunctionalProperty tells us is that all homepages 
are the homepage of _something_. Same with weblogs.

We have a similar issue with documenting foaf:mbox and its use. A common 
reply (at least earlier in FOAF's history) was "hey, what about shared 
mailboxes"? Now for foaf:mbox if I remember right we took a slightly 
different design, and said 'not all mailboxes are foaf:mbox-es, but those that
are, will be the foaf:mbox of a single Person'. With homepage, we allow more 
kinds of thing to have a homepage. 

We could have gone the other way and allowed company to have a foaf:mbox, 
while reading foaf:homepage as if it meant 'human's personal homepage'. I think
current practice shadows the way these things are used in real life, but yeah
they could certainly be more clearly documented.

This does also show I think that using OWL forces us to clean up our 
act with the documentation, as it makes inclarities clear, and confusions more
easy to spot.



* Morten Frederiksen <mof-rdf at mfd-consult.dk> [2003-06-25 00:40+0200]
> Hi there,
> danbri and I swapped opinions in #foaf regarding weblogs and homepages just 
> now [1].
> The issue is that the properties foaf:weblog and foaf:homepage are defined as 
> owl:InverseFunctionalProperties, that is, they uniquely identify a 
> foaf:Person, a foaf:Project or a foaf:Organization, just like foaf:mbox and 
> foaf:mbox_sha1sum.
> This doesn't really hold in the real world, as group and corporate weblogs 
> exist [2] [3].
> So, how to model this, and try to make sure that a group of people aren't 
> suddenly smushed into one person?
> One way would be to model groups as foaf:Project or possibly adding a 
> foaf:Group:
> [ a foaf:Person;
>   foaf:nick "mortenf";
>   foaf:interest [ a foaf:Group;
>     foaf:name "RDF IG";
>     foaf:weblog <http://rdfig.xmlhack.com/> ] ] .
> It would however be wrong to do:
> [ a foaf:Person;
>   foaf:nick "mortenf";
>   foaf:weblog <http://rdfig.xmlhack.com/> ] .
> as then everybody participating in the the RDF IG Scratchpad would be smushed 
> together.
> Any other ideas out there?
> [1] http://ilrt.org/discovery/chatlogs/foaf/2003-06-24.html#T22-11-50
> [2] http://rdfig.xmlhack.com/
> [3] http://www.sixapart.com/
> Regards,
> Morten
> _______________________________________________
> rdfweb-dev mailing list
> rdfweb-dev at vapours.rdfweb.org
> wiki: http://rdfweb.org/topic/FoafProject
> http://rdfweb.org/mailman/listinfo/rdfweb-dev

More information about the foaf-dev mailing list