RE: DCAT comments - dataset dependecy - http://www.w3.org/TR/2013/WD-vocab-dcat-20130801/

Hi Fadi,

 

Thanks to follow up.  

The suggested PAV and PROV-O and ADMS are very interesting but I could find the relationships I am looking for.

Thanks though for providing the references.

 

Focusing DCAT on the pure catalog functionality is a pragmatic objective, critical for the success of the DCAT proposal.

 

I have been looking for alternatives for the two requests I have made [1] (in subject) and [2].

 

A) For dataset dependency [1] the example use cases are the following:

-1- From the dataset catalog I find EU publications (e.g. EU legislation).

  This legislation is tagged with concepts from the EUROVOC category (but also from several other Named Authority Lists).

  The tagging typically includes a URI identifying the SKOS Concept used for the tagging.

  An application can fetch all the details of such a concept to identify the relevant concept scheme (using skos:inScheme) and then the concept scheme can be found.

  Humans can exchange information about such dependencies.

  Application can do so as well.  The question is which vocabulary to use to declare such dependencies.

  a)  PROV-O, PAV or ADMS do not provide this.  (sorry if I missed these)

  b) VoID has some definitions as suggested in [1] (and [2]) but the status and general acceptance/usage of this vocabulary is not clear to me

  Dublin Core has some potential properties dct:requires <http://dublincore.org/documents/2012/06/14/dcmi-terms/#terms-requires>  (and dct:isRequiredBy <http://dublincore.org/documents/2012/06/14/dcmi-terms/#terms-isRequiredBy> ) or dct:hasPart <http://dublincore.org/documents/2012/06/14/dcmi-terms/#terms-hasPart>  (and dct:isPartOf <http://dublincore.org/documents/2012/06/14/dcmi-terms/#terms-isPartOf> )

  (I am validating if for this use case, dct:requires is the best fit.)

-2- From the same dataset catalog I find (according FRBR <http://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records> ) I find metadata about digital objects such as “work”

  The described digital objects are further detailed going down a hierarchy of expression and manifestation.

  As in the previous use case, an application may retrieve the ‘work’ data set and then look for the specializations (expression and recursively manifestation).

  In addition to the Dublin Core suggestion form the use case -1- above, it may be an alternative to include in a ‘work’ dataset all the expression and manifestation metadata.

 

B) For the schema description (i.e. [2] - the formal specification of syntax or semantics going with a data set), I did not find an alternative yet that is better than void:vocabulary.

Such a property is useful (like but complementary to dcat:mediaType or dct:format) to know if an application can actual process the information referenced by the DCAT.

Example use case:

-1- The EU publication has it’s own owl schema.  applications not knowing this schema will have difficulty to process the data.

  Having the schema:

  i) published

  ii) referenced in the dcat:Distribution of a data set published complying to that schema

  would allow for an application to select the correct data set manifestation to be able to process or render the information.

-2- Eurostat statiscal information may be published using XML, RDF/XML, csv, html

  Especially when XML or RDF/XML is used it seems relevant to know which XML or OWL schema is used to know if the data can be processed.

 

Thanks for further advice or references.

 

[1]  DCAT comments - dataset dependecy - http://www.w3.org/TR/2013/WD-vocab-dcat-20130801/ <http://www.w3.org/mid/005901ce92bb$5b4f0cd0$11ed2670$@tenforce.com;list=public-gld-comments> 

[2]  DCAT comments - dataset schema information - http://www.w3.org/TR/2013/WD-vocab-dcat-20130801 <http://www.w3.org/mid/005001ce92b9$c61568e0$52403aa0$@tenforce.com;list=public-gld-comments> 

 

Kind Regards,

 

Johan De Smedt 

Chief Technology Officer of TenForce

 

mail: johan.de-smedt@tenforce.com

mobile: +32 477 475934

 

 

 

> -----Original Message-----

> From: Fadi Maali [mailto:fadi.maali@deri.org]

> Sent: Thursday, 15 August, 2013 11:53

> To: johan.de-smedt@tenforce.com

> Cc: vasily.bunakov@stfc.ac.uk; public-gld-comments@w3.org Comments; Alasdair J G Gray

> Subject: Re: DCAT comments - dataset dependecy - http://www.w3.org/TR/2013/WD-vocab-dcat-

> 20130801/

> 

> Hi Johan,

> 

> Thank you for your feedback on DCAT! I was wondering whether the vocabularies suggested by

> Alasdair address your needs.

> 

> Please keep in mind that DCAT is meant to be minimal and generic. Properties that are not used by

> catalogues and address specific needs were not included in DCAT. Further detailed specification can

> be defined via profiles of DCAT (see ADMS for an example  <http://www.w3.org/TR/vocab-adms/> http://www.w3.org/TR/vocab-adms/ )

> 

> Regards,

> Fadi

> --------------------------------------------------

> Fadi Maali

> PhD student @ DERI

> Irish Research Council Embark Scholarship holder

>  <http://www.deri.ie/users/fadi-maali> http://www.deri.ie/users/fadi-maali

> 

> On 7 Aug 2013, at 13:51, Alasdair J G Gray < <mailto:Alasdair.Gray@manchester.ac.uk> Alasdair.Gray@manchester.ac.uk> wrote:

> 

> >

> > On 7 Aug 2013, at 13:19, < <mailto:vasily.bunakov@stfc.ac.uk> vasily.bunakov@stfc.ac.uk> wrote:

> >

> >> I agree with Johan that  objects (datasets) dependencies are important. In the engage-project.eu,

> we have “derived” datasets published along with the “original” ones and we need to invent some

> home-made means to handle this relationship.

> >

> > Does the PROV ontology [1] capture the relationships that you are after, or perhaps you need the

> more fine-grained relationships of the provenance, authoring and versioning ontology (PAV) [2]? With

> regard to PAV they have importedFrom, retrievedFrom and derivedFrom relationships for the type of

> interaction you are describing. If you have other relationship types, I'd be very interested to hear.

> >

> > Alasdair

> >

> > [1]  <http://www.w3.org/TR/prov-o/> http://www.w3.org/TR/prov-o/

> > [2]  <http://pav-ontology.googlecode.com/svn/trunk/pav.html> http://pav-ontology.googlecode.com/svn/trunk/pav.html

> >

> >>

> >> Although despite this problem seems common, I’m not sure that a common cure for it can be

> found in DCAT. The object dependencies may be specific as the object lifecycles and the modes of

> object re-use (when the need for the expression of dependencies appears) are specific, too. So it may

> be hard to find the balance between generality and granularity to keep (most of) DCAT practitioners

> happy. Also it may contradict the DCAT metaphor of a “catalogue” that contains descriptions of

> isolated “items”.

> >>

> >> I’m not sure but some kind of “the objects/datasets dependencies ontology” may be the way

> forward – something beyond DCAT. The statements underpinned by the ontology can be then

> published in a separate registry/catalogue or shared otherwise, e.g. by making them harvestable. The

> separation of dependency statements from the objects/datasets only makes sense as you may want

> to describe the dependencies of objects/datasets registered in more than one catalogue.

> >>

> >> With kind regards,

> >> Vasily Bunakov

> >> STFC Scientific Computing

> >>

> >> From: Johan De Smedt [ <mailto:johan.de-smedt@tenforce.com> mailto:johan.de-smedt@tenforce.com]

> >> Sent: 06 August 2013 16:41

> >> To:  <mailto:public-gld-comments@w3.org> public-gld-comments@w3.org

> >> Subject: DCAT comments - dataset dependecy -  <http://www.w3.org/TR/2013/WD-vocab-dcat-20130801/> http://www.w3.org/TR/2013/WD-vocab-dcat-

 <http://www.w3.org/TR/2013/WD-vocab-dcat-20130801/> > 20130801/

> >>

> >> Dear,

> >>

> >> Based on work we are doing in the LOD2 project please consider the following.

> >>

> >> Data sets typically ma have dependencies.

> >> Examples:

> >> - subjects used for tagging a document may be detailed in one or more concept schema (available

> from the concept scheme data-set)

> >> - digital objects may have dependencies among each other (e.g. in the FRBR structure a work has

> further details in expressions; expressions have details in manifestations)

> >> Access and download of datasets may benefit (be more generic and be optimized) from having

> such dependencies formalized in DCAT.

> >>

> >> Such dependency relationship (similar to void:subset) seems not to be captured in the current

> DCAT version ( <http://www.w3.org/TR/2013/WD-vocab-dcat-20130801/> http://www.w3.org/TR/2013/WD-vocab-dcat-20130801/)

> >> If this is correct, can it be added?

> >>

> >> Kind Regards,

> >>

> >> Johan De Smedt

> >> Chief Technology Officer

> >>

> >> mail:  <mailto:johan.de-smedt@tenforce.com> johan.de-smedt@tenforce.com

> >> mobile: +32 477 475934

> >> <image001.jpg>

> >>

> >

> > Dr Alasdair J G Gray

> > Research Associate

> >  <mailto:Alasdair.Gray@manchester.ac.uk> Alasdair.Gray@manchester.ac.uk

> > +44 161 275 0145

> >

> >  <http://www.cs.man.ac.uk/~graya/> http://www.cs.man.ac.uk/~graya/

> >

> > Please consider the environment before printing this email.

> >