[rdfweb-dev] Re: <foaf:community>?

Bill Kearney wkearney99 at h...
Sun Dec 15 13:53:16 UTC 2002

> OK and we can use something like WordNet to indicate their
> status/position in the community - I had a play with this:
> <com:status rdf:about="http://xmlns.com/wordnet/1.6/Administrator">
> <foaf:Person>
> <foaf:name>Someone</foaf:name>
> <foaf:mbox_sha1sum>1239474747474747474</foaf:mbox_sha1sum>
> <rdfs:seeAlso rdf:resource="http://domain.com/foaf.rdf" />
> </foaf:Person>
> </com:status>

I'm not sure that's the right way to mark it up. You're saying the status is
about the wordnet URI. That doesn't seem quite correct. I don't have an
example handy but aren't you more interested in making a statement about the

This isn't the same but the nearestAirport idea comes to mind:
<wn:Airport air:icao="KDCA" air:iata="DCA" />

I have no idea if/how that's parseable RDF-wise. But note that it's stuffing a
wordnet element inside the nearestAirport element. I'd have to imagine some
sort of structure would be used in expressiing roles for your groups.

> but it should be possible to actually express things without that and
> go for:
> <foaf:Person>
> <foaf:name>Someone</foaf:name>
> <foaf:mbox_sha1sum>1239474747474747474</foaf:mbox_sha1sum>
> <rdfs:seeAlso rdf:resource="http://domain.com/foaf.rdf" />
> <dc:Relation>
> <rdf:Description>
> <dc:Relation.Type>isPartOf</dc:Relation.Type>
> <dc:Relation.Identifier>http://xmlns.com/wordnet/1.6/Moderator-
> 2</dc:Relation.Identifier>
> </rdf:Description>
> </dc:Relation>
> </foaf:Person>

I don't see where you're making a statement as to what group this person
moderates. I'm not sure that's the right way to use a wordnet URI as well. But
if it where you're only stating that the person is part of "something" and that
the relationship is one of being a moderator.

> and we wouldn't need to define anything for the member status
> because, as you say, being on the list would mean that you are a
> member without having to add anything else.

Well, in that the group would have to list them as a member and the person would
have to list them as being a member.

<com:member rdf:resource="http://example.com/users/own/foaf.rdf"/>

This sort of like how foaf:knows makes a statement when inside a foaf:Person.
I'm "making up" the element names here to demonstrate the structure (as in I
haven't read your group element schema yet).

> It should then be possible to use something similar to define its
> position within a group or groups with the link up to the wider group
> being a simple link while the group file would hold more data about
> the relationship with the subgroups and the communities, etc.

Sure, a group could indicate it has another group as it's members I suppose.
I'd wonder how that subgroup would indicate it "considers itself" part of that
parent group. The relationship of group having members tied to members
expressing group membership is a little easier to mentally manage. How to
escalate that up into larger group sub/super grouping gets a bit more sticky.
I'm quite concerned that it be be possible to unabiguously state 'belonging'
with some sort of provenance or trail of authority. It's one thing to say you
want to belong to something. It's another to say that someone else 'ought' to
belong to something. The user's own ability to override such external
declarations about them seems like it's going to be important. As Groucho Marx
once said "I'd belong to no group that would have me as a member."

> OK thanks for those tips - It does look like it is possible to do
> without any new namespaces. I suppose in the end the problem becomes
> that the files are increasingly bloated when a new namespace could
> simplify things and make the actual markup more meaningful.

I don't disagree that it often seems more efficient to just invent a new
namespace. And that some of the examples and first-pass examples are often
rather bloated. But it is possible to work around them and come up with
something a bit more compact while expressing useful semantics (and be
RDF-friendly in the process).

-Bill Kearney

More information about the foaf-dev mailing list