RE: Fwd: Re: Naive question on DCAT versioning

Antoine,

> 
> @Makx: the problem I'm raising assumes you've made the decision of
> creating a new version as a new resource, which is, as I understand
it,
> what ADMS would happily accommodate (see the diagram at
> http://www.w3.org/TR/vocab-adms/#asset-distribution-1).
> If you have two versions, i.e. two instances of adms:Asset, then you
> have also two instances of dcat:Dataset.
> My worry was that this "duplication" of dcat:Dataset does not really
> fit well with what I've read from DCAT. (see the body of my email at
> http://lists.w3.org/Archives/Public/public-gld-wg/2013Jul/0045.html,
> before Phil's re-wording).
> 

DCAT is completely silent on this, but I see nothing in DCAT that would
prohibit creating a new resource. 

For example, if you have a dataset that gathers weather observations for
a particular day, you will most probably have separate datasets for
different days. During a particular day, you might add observations,
say, on an hourly basis which then would create hourly updates of the
same dataset, while you would freeze one day's dataset at the end of the
day and create a new one for the next day.

DCAT has no notion of versioning, so there is no way to express (in
DCAT) that those two datasets are related -- in fact DCAT does not talk
about relations between Datasets at all. In ADMS you can use adms:next
and adms:prev between Assets that are related in this way.

Maybe DCAT needs a statement that the choice between updating an
existing Dataset or creating separate Datasets is up to the implementer.
Or should it include an example like the weather observations?


Makx.

Received on Wednesday, 31 July 2013 15:24:25 UTC