W3C home > Mailing lists > Public > public-vocabs@w3.org > December 2013

Re: Proposal for new type : Vocabulary

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Tue, 3 Dec 2013 19:20:08 +0100
Message-ID: <CAK4ZFVHdh8vPWiVGNzaM8vk+cDu3-e=U_u_gXc1xwWsvm=cX2w@mail.gmail.com>
To: Thad Guidry <thadguidry@gmail.com>
Cc: "public-vocabs@w3.org" <public-vocabs@w3.org>
Hi Thad

Thanks for the feedback

2013/12/3 Thad Guidry <thadguidry@gmail.com>

> I thought that was the whole idea behind a "namespace"... a reserved
> keyword for the vocabulary ?  (that is produced or definedBy an authority ?)
>

That's what I thought also ... ten years ago. After  almost three years of
work on the linked open vocabularies project, I can tell you that saying
vocabulary=namespace is very very far from real life.
I would say a vocabulary is a set of terms ( term = URI + label(s) +
description to make it short ) which have been packaged for some functional
purpose (glossary, classification, ontology ...)
- A vocabulary may or may not be published as a single resource under a
single URI/namespace
- The URI of the publication is not necessarily the namespace URI (number
of devilish examples available on demand)
- A vocabulary can use several namespaces, or several vocabularies can use
the same one. See e.g. http://purl.obolibrary.org/obo/ as an extreme
example. You have hundreds (thousands?) of different vocabularies
(ontologies) using this namespace


> You need both to make some sense of it all, at times.... a namespace and
> an authority who defines it and maintains the vocabulary contained within
> it.
>

Finding out the authority behind a namespace is even more tricky than
figuring out the above difference. I could forward you a very instructive
backlog of dozens of exchanges about it with vocabulary curators ... Hence
my proposal to have a simple markup to say : this is a Vocabulary, and it's
published/created/contributed by this Person/Organization, and this is a
term defined by this vocabulary. Full stop. The rest being left to
specific, hum,


> But it sounds like your asking for something concrete, "link instances"
> .... which I am trying to understand but ...ugh... "oncoming Topic
> class"... hmm..
>
> Bernard, can you give a quick example ?
>

OK, suppose I am the Library of Congress, I publish vocabularies at
id.loc.gov.
Take for example the list of MARC relators. Th html rendition of this
vocabulary is at http://id.loc.gov/vocabulary/relators.html
There are various RDF versions, but the search engines will not at the
moment browse those, right?

What you find now in the HTML is a good deal of accurate RDFa markup such as,
for the vocabulary :

<a property="rdf:type"
href="http://www.loc.gov/mads/rdf/v1#MADSScheme">MADS/RDF
MADSScheme</a>
<a property="rdf:type" href="
http://www.w3.org/2004/02/skos/core#ConceptScheme">SKOS ConceptScheme</a>

And for terms in the vocabulary :

<li rel="madsrdf:hasTopMemberOfMADSScheme skos:hasTopConcept">
<div about="http://id.loc.gov/vocabulary/relators/ard" typeof="skos:Concept
madsrdf:Authority madsrdf:Topic rdf:Property owl:ObjectProperty"> <a href="
http://id.loc.gov/vocabulary/relators/ard" property="madsrdf:authoritativeLabel
skos:prefLabel" xml:lang="en">Artistic director</a></div></li>

LoC being very liberal on types, you see that the vocabulary is both a
"MADS/RDF MADSScheme" and a "SKOS ConceptScheme", and the individual terms
have five different types, and it's linked to the vocabulary by two
different properties

Will Google and al consume this sophisticated markup? I doubt it. My
suggestion is hence to have
- schema.org/Vocabulary covering the specific types skos:ConceptScheme and
mads:MADSScheme (and elsewhere owl:Ontology, voaf:Vocabulary, foo:Glossary
bar:SubjectHeadings etc etc).
- schema.org/Topic for each term (covering skos:Concept etc).
- schema.org/isDefinedBy to formally link the latter to the former (I can
reuse in my vocabulary elements of external namespaces, and be willing to
quote their source).

And sorry about "oncoming Topic class". It means I understood that there is
a consensus about a future schema.org/Topic. Or did I miss something?

Seems to me that "Vocabulary" is as general and understandable as "Topic"
is, and schema.org would not need to go further down the jungle of various
flavours of vocabularies, see e.g., in
http://www.willpowerinfo.co.uk/glossary.htm for devilish details.

Does it clarify?


> On Tue, Dec 3, 2013 at 5:19 AM, Bernard Vatant <bernard.vatant@mondeca.com
> > wrote:
>
>> Dear all
>>
>> We're considering adding schema.org markup at lov.okfn.org, and in
>> particular in vocabulary description pages, such as
>> http://lov.okfn.org/dataset/lov/details/vocabulary_schema.html.
>>
>> We can now put each vocabulary in the broad CreativeWork type, but what
>> about a more specific "Vocabulary" type, which could be used by any kind of
>> reference vocabulary : glossaries, classifications, ontologies, concept
>> schemes, subject headings, authorities ...
>>
>> An extra would be to have a "definedBy" property to link instances of the
>> oncoming Topic class to an instance of Vocabulary.
>>
>> How does that sound?
>>
>> --
>>
>> *Bernard Vatant*
>> Vocabularies & Data Engineering
>> Tel :  + 33 (0)9 71 48 84 59
>>  Skype : bernard.vatant
>> http://google.com/+BernardVatant
>> --------------------------------------------------------
>> *Mondeca*
>> 3 cité Nollez 75018 Paris, France
>> www.mondeca.com
>> Follow us on Twitter : @mondecanews <http://twitter.com/#%21/mondecanews>
>> ----------------------------------------------------------
>>
>
>
>
> --
> -Thad
> +ThadGuidry <https://www.google.com/+ThadGuidry>
> Thad on LinkedIn <http://www.linkedin.com/in/thadguidry/>
>



-- 

*Bernard Vatant*
Vocabularies & Data Engineering
Tel :  + 33 (0)9 71 48 84 59
Skype : bernard.vatant
http://google.com/+BernardVatant
--------------------------------------------------------
*Mondeca*
3 cité Nollez 75018 Paris, France
www.mondeca.com
Follow us on Twitter : @mondecanews <http://twitter.com/#%21/mondecanews>
----------------------------------------------------------
Received on Tuesday, 3 December 2013 18:20:58 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:36 UTC