Re: Start of profiles analysis page - 2nd reply

Karen,

As far as I understand it, DCAT also makes the distinction between the
'abstract' Dataset and the 'physical' Distribution as its manifestation.

Could an Application Profile be modeled as a Dataset? Or is that a dumb
idea?

Makx

Makx

Makx.

Op 6 dec. 2017 23:07 schreef "Karen Coyle" <kcoyle@kcoyle.net>:

>
>
> On 12/6/17 10:45 AM, mail@makxdekkers.com wrote:
> > Karen,
> >
> >> Not all access to APs will be through content negotiation, AFAIK, so we
> have to consider
> >> other access avenues, such as a document at is located on a web site,
> profiles in wikis, etc.
> >
> > The expressions of the profile might be at
> >
> > http://example.org/profiles/xyz/profile.rdf
> > http://example.org/profiles/xyz/profile.xml
> > http://example.org/profiles/xyz/profile.json
> >
> > So it would be possible to access them without content negotiation. But
> I guess, we need to consider content negotiation because our deliverable is
> called "Content Negotiation by Application Profile"
> >
> >> If there is a "concept" AP it needs to be something that can be
> represented,
> >> thus is not entirely abstract.
> >
> > In my mind, it *is* "abstract" in the same sense that FRBR Work is an
> abstract entity.
>
> Makx, the FRBR work is proving to be very difficult to implement
> precisely because it is so hard to be precise about an abstraction. If
> the AP is "abstract" in that sense it has no actual existence in any
> written or coded form, which means that it cannot be "converted" to rdf,
> html, xml, or whatever. It is ethereal, an essentially non-existent as
> any "thing". I don't know how we can work with such an entity.
>
> kc
>
>
> >
> > Makx.
> >
> >
> >
>
> --
> Karen Coyle
> kcoyle@kcoyle.net http://kcoyle.net
> m: 1-510-435-8234 (Signal)
> skype: kcoylenet/+1-510-984-3600
>
>