Re: Media Types URIs

Hi Nick,


> The comment for dct:format [1] reads “Examples of dimensions include size
> and duration. Recommended best practice is to use a controlled vocabulary
> such as the list of Internet Media Types [MIME]” so surely use of these
> URIs with it is appropriate?
>

Yes, this is stated in dcterms. And, since dcat:mediaType is
rdfs:subPropertyOf dcterms:format, from the point of view of dcterms this
is fine (values are formats and, among them, media types).


> Perhaps more appropriate for general work than another list of file types
> due to the IANA list being very well known and widely used but not
> appropriate if specific file types are needed but not in the IANA list. The
> DCAT-AP list doesn’t look long, perhaps all its entries are in the IANA
> list? I could check this.
>

I totally support using the IANA list. What I wanted to point out is that
DCAT-AP v1.1 (https://joinup.ec.europa.eu/release/dcat-ap-v11) refines the
description of dcterms:format and dcat:mediaType, specifying that
dcterms:format has to have values from the EU file type list, and
dcat:mediaType values should be IRIs from IANA such as
https://www.iana.org/assignments/media-types/text/turtle. The two lists are
complementary. They have some overlap, but the both contain formats the
other one does not. For instance GMZ (zipped GML files) has an entry in the
EU File Type list, but it does not have a media type. In fact, some sources
say that application/x-gmz should be used, but that is not an IANA
registered media type.


> For dcat:mediaType (not dcat:mimeType!), I see that indeed the comment
> reads “This property SHOULD be used when the media type of the distribution
> is defined in IANA, otherwise dct:format MAY be used with different values”
> [2] so it would be appropriate and suggested to use these W3IDs for
> dcat:mediaType but I would add that, where possible if using dct:format,
> these W3ID Media Types too. In the ProfileDesc Ontology, as in many other
> places, dct:format is used and dcat:mediaType is not used so there I’ll be
> using these values.
>

Indeed, I meant dcat:mediaType, sorry. Since dcat:mediaType is subproperty
of dcterms:format, I would say the comment means "if it has a media type,
use dcat:mediaType. If it does not have a media type, use dcterms:format
with something else". At the same time, a reasoner can interpret using a
media type with dcat:mediaType as using the media type with dcterms:format
thanks to the subproperty.

Again, DCAT-AP refines this with the two codelists clearly specified. I
realize that this does not have to be reflected back in DCAT, but maybe it
should. I don't see the value of using the same media type explicitly both
in dct:format and dcat:mediaType.


> Note that in the DCAT documentation, dcat:mediaType ise used like this:
>
>
>
> dcat:mediaType "text/csv" ;
>

I wonder if this is an error in DCAT 2014 examples since the range of
dcat:mediaType is dcterms:MediaTypeOrExtent, which is a rdfs:Class, not a
rdfs:Literal.

>
>
> And I would prefer:
>
>
>
> dcat:mediaType <https://w3id.org/mediatype/text/csv> ;
>
> This creates inconsistency with DCAT-AP where it should be
https://www.iana.org/assignments/media-types/text/turtle, even though this
is not dereferencable.


> Regarding official use in DCAT: well let’s see! I like the idea of using
> most basic/widely understood and most widely accessible codelists so
> perhaps this list of Media Types now as LD is a challenge to the EU to step
> up their list accessibility!
>

I also like this idea. The EU lists are available as RDF dumps, and now I
see that at least the code list IRIs are dereferencable, if not the
codelist items.

Best,

Jakub


>
> [1] http://www.dublincore.org/documents/dcmi-terms/#terms-format
>
> [2] https://www.w3.org/ns/dcat#
>
>
>
>
>
> *From: *Jakub Klímek <jakub@jakubklimek.com>
> *Date: *Tuesday, 19 June 2018 at 5:17 am
> *To: *Nicholas Car <Nicholas.Car@csiro.au>
> *Cc: *"public-dxwg-wg@w3.org" <public-dxwg-wg@w3.org>
> *Subject: *Re: Media Types URIs
>
>
>
> Dear Nick,
>
>
>
> Are you suggesting using your IRIs such as [1] for MIME-Types with
> dct:format?
>
>
>
> In DCAT-AP, dct:format is actually used with the EU File Type codelist [2]
> concepts such as [3], whereas for IANA based MIME Type IRIs such as [4],
> dcat:mimeType is used. Indeed, neither the File Type, nor the MIME Type
> IRIs are dereferencable. Still, some consistency would be nice.
>
>
>
> BTW I have been asking the EU publications office to make the EU MDR NALs
> (now EU vocabularies) dereferencable for 3 years now, unfortunately with no
> success. Nevertheless, their use is mandatory in DCAT-AP, which could be
> reflected in DCAT since there are no other official code lists for those
> properties.
>
>
>
> Best regards,
>
>
>
> Jakub Klímek
>
>
>
> [1] https://w3id.org/mediatype/text/turtle
>
> [2]
> https://publications.europa.eu/en/web/eu-vocabularies/at-dataset/-/resource/dataset/file-type
>
> [3] http://publications.europa.eu/resource/authority/file-type/RDF_TURTLE
>
> [4] https://www.iana.org/assignments/media-types/text/turtle
>
>
>
> On Mon, Jun 18, 2018 at 7:42 PM Car, Nicholas (L&W, Dutton Park)
> <Nicholas.Car@csiro.au> wrote:
>
> Dear DXWG,
>
>
>
> I’ve just put up a small API to deliver Media Types via URIs so if you
> want to quote text/html, you can do so like this:
> https://w3id.org/mediatype/text/turtle and get a simple web page or this:
>
>
>
> @prefix dct: <http://purl.org/dc/terms/> .
>
> @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#
> <http://www.w3.org/2000/01/rdf-schema>> .
>
> @prefix xsd: <http://www.w3.org/2001/XMLSchema#
> <http://www.w3.org/2001/XMLSchema>> .
>
>
>
> <https://w3id.org/mediatype/text/turtle> a dct:FileFormat ;
>
>     rdfs:label "turtle"^^xsd:string ;
>
>     dct:contributor <https://w3id.org/mediatype/Eric_Prudhommeaux>,
>
>         <https://w3id.org/mediatype/W3C> .
>
>
>
> This just replaces a system that did this using purl.org/NET/mediatypes/
> but which has been down for years (see
> https://gist.github.com/stain/4635250).
>
>
>
> The reason for doing this now is that we are really using dct:format a
> lot but don’t have dereferenceable URIs for it.
>
> The register of Media Types is here: https://w3id.org/mediatype/
>
>
>
> Consider this tool actively developed so you can raise issues (
> https://github.com/nicholascar/mediatypes-dataset/issues) or get things
> changed if they aren’t great!
>
>
>
> Cheers,
>
>
>
> Nick
>
>
>
>
>
>