[rdfweb-dev] Time's running out?

Jim Ley jim at jibbering.com
Thu Aug 7 22:38:55 UTC 2003


"Jonathan Greensted" <jonathan.greensted at sentient.co.uk>
> Hello FOAFers
>
> Is Nick correct?
>
> http://www.ecademy.com/module.php?mod=club&op=forum&c=628&t=21142#endm

He makes some interesting points, but has failed to understand RDF, making
up extra namespaces isn't dangerous, the fact that I use foaf:wakeUpSlacker,
and it's not written in the spec isn't dangerous, it just means people will
get a triple with no defined meaning - other than the fact I'm not the
controller of the xmlns.com URI space so shouldn't really place any
semantics on uris in it - there's nothing wrong with creating extra ones, it
doesn't break anything other than the fact that different people can create
the same name with different semantics, the terms found in our crawls which
aren't defined in the foaf RDF document (which doesn't mean they don't
exist, since there's no requirement in RDF to have a machine readable
document which defines the terms, that's just best practice.) have generally
been used as agreed experiments in the foaf space by foaf developers.

Other RDF namespaces coming along to be used to define almost FOAF terms
would be good news, not bad news, if FOAF doesn't map to the outlook model
that's fine, the outlook model can be defined in a different RDF vocab,
preferably re-using terms from other RDF vocabs where possible (since it's
always best not to invent your own)  for example the mobileNumber etc. can
be easily modeled just with a  fredsRDF:mobileNumber - you can then use some
RDF saying it's a subProperty of foaf:tel and that is what the power of RDF
gives us. No single vocabulary will ever deal with all use cases, if you try
and mould foaf to be outlook export specific it won't deal with other
situations.

"Sinking under a swamp of namespaces" simply won't happen, RDF doesn't care
how many namespaces (indeed it knows little about namespaces at all, they're
an effect of the RDF/XML syntax) you have in it, it's only the folks still
looking at it from an XML format.

Victor Lindesay's comment that " the list above which must be an
embarrassment to both the owners of this so called standard and all the
major participants " I don't agree with at all, I'm not an owner, and I'm
not sure if I'm a "major participant" but I don't find it embarrassing,
they've been known and discussed for ages, both me and Libby have published
the same stats for a long time.

Again I feel it's someone who doesn't fully understand RDF vocabularies, and
thinks in terms of XML validation, and thinks in terms of having one RDF
vocab that does everything, the task is not to create _an_ RDF schema for
representing a person and people they know, the task is to create RDF
schemas which achieve it, it will never be possible to create one schema
that does this.

Jim.




More information about the foaf-dev mailing list