W3C home > Mailing lists > Public > public-vocabs@w3.org > April 2012

Re: Schema.org External Enumerations mechanism

From: jean delahousse KC <jean.delahousse@knowledgeconsult.com>
Date: Tue, 24 Apr 2012 17:34:43 +0200
Message-ID: <CAO+52yU=RhZfE1N9E97iL6cmQJVt4c4AsOQKZa1jyjVzd4+2Ow@mail.gmail.com>
To: Martin Hepp <martin.hepp@ebusiness-unibw.org>
Cc: Guha <guha@google.com>, Danny Ayers <danny.ayers@gmail.com>, Dan Brickley <danbri@danbri.org>, public-vocabs@w3.org
Hi all,

I have been reading the exchanges about External Enumerations mechanism and
the proposal.

I have few remarks mainly based on schema.org mission and on political
issues :

1) it is a very different mission to propose and maintain a schema to
describe web pages content than to select and promote authorities.
Does this new huge task should be endorsed or asked by Schema.org ?

2) if schema.org makes a real work to qualify and select authorities it
would mean to have organized teams able to evaluate authorities in multiple
knowledge domains, in various languages, for various points of view and
even so the work cannot be done. An other solution would be for Schema.org
to only select the "well-known", "widely used" authorities, which can be
very tricky in some specialized domain (health for example) and create a
lot of conflict between authorities editors.

3) if schema.org does not do a work of qualification and selection of
authorities, but only publish community suggestions, it means every group
or person who wants to promote his point of view on a domain will reference
its authority on any domain and in any language without any quality
control. This would only weaken Schema.org position.

The enrichment and maintenance of schema.org schema already seems a huge
and long term task. Schema.org's work on a schema to describe web page
content can already be interpreted as a Google control on knowledge
representation. Going a step further and give an advice on authorities
which reflect the diversity of points views on knowledge domains can become
quite a political issue, especially in a world where SEO people make the
final decision in the media and publishing houses.

The selection of the best authorities in the various domains, cultures and
languages may go on as an open, creative and very interesting process where
the various authority editing bodies, including wikepedia, freebase, but
also russian, chinese, egyptian, french, brezilian... organizations...  ...
compete and collaborate to publish their authorities and interconnect them,
letting some most popular URI emerge and be inter connected to others...
with a schema.org most neutral and friendly attitude.

Best wishes

Jean Delahousse


2012/4/23 Martin Hepp <martin.hepp@ebusiness-unibw.org>

> Hi Guha, all:
>
> Note that for using Wikipedia URIs as identifiers for *types*, you can
> readily use
>
>    http://www.productontology.org
>
> It provides class definitions including translations for any lemma in
> Wikipedia. This could also be more strongly aligned with schema.org. It's
> also very scalable, running on Google App Engine.
>
> For *individuals*, I would suggest to recommend DBpedia URIs but be open
> to other instances defined in the local namespace of a single Website. We
> use the same pattern e.g. in
>
>    http://purl.org/vso/ns#feature
>
> For a rationale for www.productontology.org, see
>
>    http://www.productontology.org/#faq
>
>
> Best wishes
>
> Martin Hepp
>
> On Apr 23, 2012, at 7:09 AM, Guha wrote:
>
> > Your example is really good and has got me rethinking ...
> >
> > Let me think about it a bit more and get back to you.
> >
> > guha
> >
> > On Sun, Apr 22, 2012 at 9:49 AM, Danny Ayers <danny.ayers@gmail.com>
> wrote:
> > On 22 April 2012 15:41, Danny Ayers <danny.ayers@gmail.com> wrote:
> >
> > Just to clarify with an example...
> >
> > > But where is the tangible utility in using schema.org URLs? As noted
> > > earlier, they actually add friction to the system.
> >
> > ...ok, I'm visiting my mother this week, and a couple miles down the
> > road is a hill fort called "Fin Cop". So how would I talk about that
> > in microdata?
> >
> > I've no idea offhand what detail is available in schema.org for
> > classifying places, but I can remember there is some coverage. So I
> > start at:
> >
> > http://schema.org/Place
> >
> > I see:
> >
> > http://schema.org/LandmarksOrHistoricalBuildings
> >
> > and using an example from the Place page I already have:
> >
> >  <div itemprop="location" itemscope
> > itemtype="http://schema.org/LandmarksOrHistoricalBuildings">
> >    <a itemprop="url" href="...to be decided...">
> >    Fin Cop
> >    </a>
> >  </div>
> >
> > But which URL to use?
> >
> > Wikipedia is blessed, so I search there and find:
> >
> > http://en.wikipedia.org/wiki/Fin_Cop
> >
> > Paste that into my markup, job done.
> >
> > Unless I want to use the schema.org alias. In which case I have to
> > look up the appropriate template/mapping, apply it, and then use that
> > URL.
> >
> > http://ext.schema.org/wikipedia/en/Fin_Cop
> >
> > Job done - after an extra step.
> >
> > Out of curiosity I had a quick go at getting a term for describing a
> > place in a similar fashion using existing RDF vocabs. Starting with:
> > http://sindice.com/search?q=Place
> > a couple of clicks later I had:
> >
> > http://sw.opencyc.org/2009/04/07/concept/en/AncientSite
> > Same as:
> > http://umbel.org/umbel/sc/AncientSite
> >
> > - though there appear to be a lot of other alternatives.
> >
> > Putting "Fin Cop" into Google search, the most compelling-looking URL
> > for the place on the first page of hits is the Wikipedia one (4th on
> > the list here).
> >
> > So as far as the effort needed to find suitable terms, there wasn't
> > really very much to choose between them. For data consumers, it seems
> > probable that in due course the schema.org class will be more useful
> > simply because of wider deployment. But well-known vocabularies like
> > http://en.wikipedia.org/wiki/* are already widely deployed on the web
> > (as regular links)...so why bother aliasing them?
> >
> > Cheers,
> > Danny.
> >
> > --
> > http://dannyayers.com
> >
> > http://webbeep.it  - text to tones and back again
> >
>
> --------------------------------------------------------
> martin hepp
> e-business & web science research group
> universitaet der bundeswehr muenchen
>
> e-mail:  hepp@ebusiness-unibw.org
> phone:   +49-(0)89-6004-4217
> fax:     +49-(0)89-6004-4620
> www:     http://www.unibw.de/ebusiness/ (group)
>         http://www.heppnetz.de/ (personal)
> skype:   mfhepp
> twitter: mfhepp
>
> Check out GoodRelations for E-Commerce on the Web of Linked Data!
> =================================================================
> * Project Main Page: http://purl.org/goodrelations/
>
>
>
>
>


-- 
_____________________________________________________________________
Directeur Associť, *KnowledgeConsult  *http://knowledgeconsult.com
jean.delahousse@knowledgeconsult.com  +33 (0)6-01-22-48-55 skype:
jean.delahousse  * *twitter.com/jdelahousse
Received on Tuesday, 24 April 2012 15:35:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:49:02 GMT