W3C home > Mailing lists > Public > public-vocabs@w3.org > November 2011

Re: FAQ : Is this vocabulary currently curated?

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Tue, 22 Nov 2011 15:08:42 +0100
Message-ID: <CAK4ZFVEdYMqwcHRuK4ROw-74b080Ffz8CZjP-29Qfy2EFECxoA@mail.gmail.com>
To: Martin Hepp <martin.hepp@ebusiness-unibw.org>
Cc: public-vocabs@w3.org
Hi Martin

I think basically I agree with all the points you make. The (SW or other)
vocabulary publishers being what they are, it is certainly useless to go on
saying they should do this or that.
But I like your idea of proper dead ontologies disposal. As we well know,
there are basically three ways of disposal : trash anywhere by any means,
bury and forget about it, or sustainable recycling. The two first options
are actually what we've seen so far. Unattended litter, or 404. Recycling
is something yet to explore. Having some sort of repository of abandoned
URIs with possible equivalents in living vocabularies would be a good idea
- at least for those URIs unfortunately embedded in legacy datasets. This
is a track I'm willing to explore further.

Bernard

2011/11/22 Martin Hepp <martin.hepp@ebusiness-unibw.org>

> Hi Bernard,
> On Nov 22, 2011, at 10:00 AM, Bernard Vatant wrote:
>
> > Hi Martin
> >
> > Thanks for breaking the silence, even to say that I should not waste
> time and energy on this :)
> >
> You are welcome! I did not want to say that the approach is a sure waste
> of time; I just wanted to explain why it is  low priority for me as an
> ontology maintainer.
>
> > 2011/11/21 Martin Hepp <martin.hepp@ebusiness-unibw.org>
> > Hi Bernard:
> >
> > My position is that while such meta-data can be useful, it is not
> mission-critical, because it is uncertain whether a declarative approach is
> sufficient or appropriate in here.
> >
> > Not sufficient certainly but useful nevertheless if it allows any
> potential adopter to send a signal to the actual curator and get feedback.
> It's not simply putting some declarative boolean (curated Y/N) in the
> vocabulary, but ensure that if I have questions or suggestions about the
> vocabulary evolution, there is a place I can send it and I'll get feedback.
> >
> > I would assume that in real, Web-scale scenarios, you need to derive the
> respective information from a multiplicity of signals, e.g.
> >
> > - adoption rate in data
> >
> > Adoption rate is a signal that the vocabulary is useful, not that there
> is still someone taking care of it. Take the example of
> http://www.aktors.org/ontology/portal. This vocabulary is the kind of
> "academic" ontology you like so much. The last publication date is 2003.
> The documentation page indicated in dc:source and rdfs:seeAlso
> http://d3e.open.ac.uk/akt/2002/ref-onto.html seems dead, making your
> point that it has been developed in a research project and forgotten on a
> shelf.
> > It's used nevertheless in quite a few datasets. See
> http://thedatahub.org/tag/format-akt. Of course you may argue that those
> datasets have as little significance as the vocabularies they stand upon,
> but ...
> >
>
> There may be cases like that, but I would assume that on a global scale,
> people will not keep up support for a format that is no longer supported by
> its owner, at least as soon as someone has to justify the underlying
> investment in a corporation. A good way to handle such cases would be for
> someone to republish the ontology under a new base URI and to add sameAs,
> equivalentClass, equivalentProperty axioms.
>
>
> > - HTTP header information
> >
> > Not sure how HTTP information gives you clue about curation.
> For example
>
> Last-Modified: Fri, 18 Nov 2011 07:36:54 GMT
> Content-Length: 186700
> and maybe
> ETag: "11080169-2d94c-4b1fd66a87180"
>
>
> > etc.
> >
> > I would expect that most dead vocabularies on the Web, created in some
> PhD thesis project and long abandoned by its creator would still bear the
> "actively maintained" flag in the code.
> >
> > Indeed, has 404 documentation as seen above. But not curated does not
> mean "dead" if they are indeed used.
> Yes, but I think as soon as adopting the ontology has a business
> dimension, I think that a non-maintained ontology will be the very last
> resort.
>
> >
> > Also, as of now, the number of relevant Web ontologies is pretty small,
> >
> > I'm no-one to decide if an ontology is "relevant" or not. You seem to
> know best
> I am just frank in my assessment of which ontologies will fly at Web scale.
>
> > and have decide that anything coming out of a PhD thesis is irrelevant.
>
> I never said that. I am just saying that developing Web ontologies that
> are here to stay is a huge effort, while important for the SW to fly, and
> that I am puzzled by the limited willingness of the SW community to build
> the ontologies needed for realizing the vision they perpetually preach.
>
> > All users of the "aktors" ontology above have found it useful at some
> point, I'm no one to judge if they were right to do so or not.
> Me neither, but again - it would be better to take care of the good work,
> deploy it in a new namespace, and make sure it is properly maintained.
>
> Abandoned ontologies are similar to unexploded ordnance devices from WW
> II. They put uninformed visitors to the SW grounds at risk, and some day, a
> brave specialist has to disarm them to save all the data.
> It would be way better if the community properly disposed of their
> abandoned ontologies from the SW 1.0 era.
>
> > so it is unclear whether an automated approach is cost-efficient.
> >
> > I don't know what you mean by that.
>
> I meant that you cannot naively interpret the signals from your
> annotations anyway, so the question is whether the total effort of getting
> this approach adopted is justified.
> Your proposal will only make it if you are willing to fight for it for the
> next years, provide support and assistance, foster tool development, etc.
>
> No Web ontology or extension will fly if you just propose it to the
> public. An specification without community support is not yet an ontology
> ;-)
>
>
> Martin
>
>
>


-- 
*Bernard Vatant
*
Vocabularies & Data Engineering
Tel :  + 33 (0)9 71 48 84 59
Skype : bernard.vatant
Linked Open Vocabularies <http://labs.mondeca.com/dataset/lov>

--------------------------------------------------------
*Mondeca**          **                   *
3 cité Nollez 75018 Paris, France
www.mondeca.com
Follow us on Twitter : @mondecanews <http://twitter.com/#%21/mondecanews>
Received on Tuesday, 22 November 2011 14:09:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:57 GMT