[rdfweb-dev] foaf:knows foaf:knowsWell foaf:friend ... : time to choose!

Seth Russell seth at r...
Mon Mar 12 19:55:11 UTC 2001

Dan Brickley:

we have foaf:knows foaf:knowsWell foaf:friend all jumbled up. 'friend'
was the original, but I took it out and put knows/knowsWell instead, but
never gave them a good definition.

I want this to be easily adopted, and make intuitive sense to people,
and not make them delay their rdf self-description by forcing them to
classify everyone they know as 'knows well' or (merely)'knows'.

current thinking: flatten the core 'knows' down to a single loose
relation 'foaf:knows'; no 'friend' or 'knows well' stuff. Instead, we
let them tell the Web their story via other factoids (event logs,
co-authorship, working for the same organisation, appearing
in same holiday pictures... whatever). But keep the core of FOAF to


Sounds great to me. But may I suggest foaf:knowsOf (of perhaps even
foaf:acquaintedWith) rather than foaf:knows. To me the word "knows" is
far to general to be used for this purpose. True, we can prefix it with
foaf and that makes it specific, but who is going to remember that when we
have thousands of *common* words floating around. So, for this common word
you are forcing people to remember that it was coined in the namespace
prefixedWith "foaf" ... guys like me with poor memories are fucked. Right
now I can't remember which words belong in rdf, rdfs or dmal .. there is no
rhyme or reason to their placements there ... luckly there are fue if any
overlaps. But when we have hundreds of schema in popular usage (some of
them overlapping) how can we remember? What do you think of a kindOf
agreement that common words that are coined for use by the Semantic Web as
specific arc labels have some simple rules for formation, are not spelled
exactly like common English words, and that the schema makers try very hard
not to use words that have already been used ?

Another thing is that someone (perhaps even me) is going to come along and
declare the english prefix and then suddenly all English words can be used
as arc labels ... so that then we will have english:knows .. which carries
with it all the semantics and beautiful ambiguity given to it by the
English language. Keeping the more precise words of a schema apart from
these words, methinks will be a big advantage down the road a couple of

... how does that sound?


More information about the foaf-dev mailing list