Eurovoc License ?

02/04/2013

The DCAT profile mentions Eurovoc as proposed vocabulary, which is an excellent suggestion.

However, the SKOS/XML version of Eurovoc still seems to require that a (free) license is obtained from the Publication Office. It would be nice if this licensing step is removed, and Eurovoc simply becomes available as open data under a Creative Commons License.

(IIRC e.g. the European environment thesaurus GEMET is available under CC-BY license)

Component

Documentation

Category

feature

Comments

Tue, 02/04/2013 - 10:44

Hi,

We adopted this recommandation in our DCAT implementation through the CKAn software but we had to ask first for a agrement from the publication office.  As a result this is not possible to link to the SKOS concept as a dereferencable URL thus the interoperable added value is limited.

As a result we have the top SKOS concept in the dcat:themeTaxonomy tag <dcat:themeTaxonomy>Environnement</dcat:themeTaxonomy> and the indexation concept in the dcat:theme tag<dcat:theme>Recyclage des déchets</dcat:theme> but I would like to be able to link to the eurovoc.eu url addind the multilingual aspect of the Thesaurus

 

Pascal
 

Tue, 02/04/2013 - 11:13

This is a very good point. We can discuss this under agenda item 8 in tomorrow's virtual meeting.

Makx.

Wed, 03/04/2013 - 13:11

EUROVOC is perfect as a common thesaurus, but the WG shoud consider the possibility of supporting also other ones, especially those already used in existing metadata and possibly available according to Linked Data principles. A typical example is GEMET [1], already mentioned by Bart when opening this issue and extensively used in the environmental sector.

As far as INSPIRE metadata are concerned, the regulation [2] requires what follows:

If a resource is a spatial data set or spatial data set series, at least one keyword shall be provided from the general environmental multilingual thesaurus (GEMET) describing the relevant spatial data theme as defined in Annex I, II or III to Directive 2007/2/EC.

Such thesaurus is hosted by GEMET [3], but it is also available as Linked Data from the INSPIRE Feature Concept Dictionary [4].

Another issue is about the adoption of EUROVOC in metadata that originally do not make use of it. Different options are available to deal with this, and the suitability of each of them may vary depending on the metadata provider. For instance, the creation of mappings between EUROVOC and the thesauri actually used in the existing metadata is probably the most effective approach, but, whenever it cannot be readily implemented, alternative & temporary solutions can be adopted. The WG may consider preparing guidelines and best practices for metadata providers on how to deal with this issue.

  1. http://www.eionet.europa.eu/gemet
  2. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32008R1205:EN:NOT
  3. http://www.eionet.europa.eu/gemet/inspire_themes
  4. https://inspire-registry.jrc.ec.europa.eu/registers/FCD

Wed, 03/04/2013 - 18:46

Good point Andrea.

I suggest to add some text in section 9 that says that for the properties listed, a data provider should at least use a term from the controlled vocabulary that is listed, but may choose to provide terms from other vocabularies as well. 

The idea of guidelines and best practice for mapping from local vocabularies to common ones is a good idea. Maybe there are WG members who have experience or can even link to existing guidelines or tools?

Thu, 04/04/2013 - 19:07

This is an important topic. Maybe the document should reflect that any vocabulary used should:

  1. Be published openly under an open license
  2. Be dereferencable

Without these requirements it will be difficult for an aggregator to support the use case of increasing findability by clustering datasets by topic.

Also, if different vocabularies are used it is likely that mapping will never occur. Mapping controlled vocabs can be tedious so if one catalog has used a local vocabulary it may be too much work to map it to Eurovoc.

Tue, 09/04/2013 - 06:41

The Publications Office is taking all necessary steps to modify the licence policy that currently regulates EuroVoc, so as to make it accessible without the need to ask for such agreement any longer. The decision will be formalised by the Management Committee in May

Tue, 16/04/2013 - 10:23

+1 to Peter's propossal. Any propossed reference taxonomy or vocabulary (being finallly EUROVOC or any other) should be available under an open license. But, given that in fact we are promoting the use of Semantic Web technologies by adopting DCAT as a reference standard, it doesn't make too much sense that we may encourage usage of any vocabulary that it is not dereferencable. Both should be minimal requirements for any proposed vocabulary, although there may be also others.

Wed, 17/04/2013 - 09:37

We can try to come up with some basic characteristics of controlled vocabularies.

In addition to:

  • open licence to enable reuse
  • dereferenceable to enable 'follow-your-nose'

Possible other criteria:

  • described using well-known predicate vocabularies (SKOS, DC, FOAF, DCAT etc.)
  • explicit maintenance commitment by an organisation
  • explicit persistence policy

 

Wed, 17/04/2013 - 22:13

See also this thread by Carlos Iglesias on requirements for controlled vocabularies:

https://joinup.ec.europa.eu/discussion/requirements-controlled-vocabularies

Wed, 26/03/2014 - 21:04

From today you no longer need to register to download the SKOS and XML version of EuroVoc: http://eurovoc.europa.eu/drupal. By clicking on the download link you will be redirected to the ODP website from where you can download the EuroVoc resources. You can find the same resources as well on the Metadata Registry website: http://publications.europa.eu/mdr/eurovoc/index.html.

Thu, 27/03/2014 - 16:56

Ah, that's great news, thanks.

Login or create an account to comment.