W3C home > Mailing lists > Public > public-lod@w3.org > January 2011

Fwd: Vocabulary of a Friend (VOAF)

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Thu, 20 Jan 2011 01:10:45 +0100
Message-ID: <AANLkTi=Y9cOSY8KpfpXOdxkqi+fT5XGYAYob_PDgCiTg@mail.gmail.com>
To: Linking Open Data <public-lod@w3.org>
Hi Dan

cc to LOD as suggested ...

It can help developers get an overview of the jumble of RDF
> vocabularies, and understand how eg. SIOC, FOAF, DOAP, SKOS, Dublin
> Core, etc are interconnected. It can help vocabulary managers know
> what else is out there, and understand the potential impact of changes
> to term definitions. And it can help present the idea of the Semantic
> Web as a distributed, decentralised and collaborative system - too
> often it is seen as the opposite, "one big ontology" etc.

This and many more. I would like to add a terminological interlinking based
on lexvo.org terms, such as

<http://lexvo.org/id/term/eng/Person>  lvont:means  <

I just saw the LOD mailing list thread too. Glad to see it announced!

And tweeted about ... Twitter / Search -

If X voaf:reliesOn Y, and Y voaf:reliesOn Z, does this imply that
> X voaf:reliesOn Z?

I would not say that, or it depends on the flavour of "reliesOn". It would
be true in the case where "reliesOn" means extension of class or property,
or attaching a property of X on a class of Y etc. But if Y is used to add
metadata to X, it's another story.
So maybe it would be a good idea to define two different subproperties,
"uses" and "extends".

 Of course, it's all fine and I'm flattered to have FOAF used this way.

Thanks! Maybe you don't deserve it, but FOAF does :)

> In fact, I was long uncomfortable with the early LOD diagrams, since
> they represented FOAF as a single data source; I rather see it as a
> vocabulary, or as a project that has a vocabulary. So this
> characterisation is more accurate and useful, I'm sure.


> The 'reliesOn' relation is useful as a high-level overview, but I
> think would perhaps be more useful alongside an aggregation in SPARQL
> of all the relevant schemas.

Not sure I understand what you mean here ...

> I have experimented occasionally with
> this, but never built a solid service. The idea would be to have at
> least the current DC, FOAF, SKOS etc etc schemas in a quad store (each
> in a separate named graph). Ideally I'd also like historical versions,
> so we could see the evolution of the vocabulary network over time.


> But the main thing would be to be able to run queries such as 'find me all
> the subclasses of dct:Agent' or 'all the properties whose range is
> (directly or indirectly) foaf:Document'. I'm just starting to look at
> this again for another project, hopefully I'll have a discussion
> document on that available soon...

Oh yes that's another story, but interesting as well.
The initial aim of VOAF was to keep at vocabulary metadata level, not to
drill down into elements, but ...

How do you plan to proceed with mapping out 'reliesOn'? I wonder how
> you can distinguish an assertion in a schema that is there really as a
> public service convenience mapping, rather than a strong dependency?
> There is a sense in which SIOC's classes for example 'rely on' FOAF,
> since they were originally defined as extending FOAF, and in some
> sense inspired by it.


> However if I put reciprocal assertions into
> FOAF's RDFS, does this mean we 'rely on' SIOC?

Good point. In a sense I would say yes. You (FOAF as a vocabulary and
community of users) rely on the fact that SIOC relies on you, so to speak.
IOW, I trust the fact that you trust me. Like in human relations, actually.

> Figuring out where
> there is a strong dependency (rather than a case where projects are
> documenting the overlapping meaning of their terms) is perhaps
> something more at the social level between projects. Would it make
> sense to think of VOAF as a way for projects to be more explicit and
> say which other projects they 'rely on', and for what? ie. that it is
> a tool for vocabulary managers to express things that we couldn't
> otherwise know.

I think along those lines. I gave a tutorial yesterday at SemWebPro in
Paris, wherein I insisted that a good (useful) vocabulary means a community
of users and a responding publisher. Otherwise it's a dead vocabulary. Might
be as smart as can be, but dead altogether. And the semantic network of
vocabularies is backed by a social network. That's why I liked so much to
call it VOAF. I think we could easily mirror the network of vocabularies by
foaf:knows network linking their creators.

> As an example, there is the business of foaf:Agent and dct:Agent being
> mapped. How would 3rd parties know whether to say they 'rely on' each
> other? It's interesting that you're doing this (if I understand right)
> at both the vocabulary and term level.

I don't do it by now, but vaguely intend to do so, as said above, using the
lexvo.org terms glue. Strangely enough, we lack a predicate to assert that a
class or property belongs to a vocabulary. The equivalent of skos:inScheme.
What would you use? dcterms:partOf ? In any case it's a more ambitious task.

> I certainly find it easier to say that FOAF (as a vocabulary / project)
> "relies on" the Dublin Core
> community's vocabulary to provide detailed descriptions of documents
> and bibliographic content, or that it "relies on" SIOC when there's a
> need to describe eg. forums or bulletin boards.

Hmm. Interesting. Introducing facets or contexts in which "relies on"
applies. I have to munch over this.

> Expressing 'relies'
> links between terms is harder. I like to add mappings; but I don't
> like to add dependencies. So my guess is VOAF will be easier to use as
> a kind of 'vocabulary buddylist' than at the term level, and for
> terms, we might turn directly to things like subClassOf /
> subPropertyOf...

Indeed. But the terminological glue is something to think about.

> ps. feel  free to migrate this to public-lod

Done :)


Bernard Vatant
Senior Consultant
Vocabulary & Data Engineering
Tel:       +33 (0) 971 488 459
Mail:     bernard.vatant@mondeca.com
3, cité Nollez 75018 Paris France
Web:    http://www.mondeca.com
Blog:    http://mondeca.wordpress.com
Received on Thursday, 20 January 2011 00:11:15 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:16:11 UTC