RE: Thought for URI section

The Krems Share-PSI meeting should be right in the middle of the public review period of the DCAT-AP so it would be good to talk about the evolution of DCAT based on the experience with the Application Profile over the last two years.

I would be interested in your view about how to model the time series example that you give in DCAT. After all, we want DCAT to be the preferred standard for data on the Web, but examples of how a human user gets to a landing page for a dataset (which is what http://index.okfn.org/place/ is) do not quite match the way DCAT models a dataset and its distributions.

In a way, having a latest version URI may be contrary to some cool URI principles: in your example (1) the URI for the 2015 data will change; in January 2016 it will no longer be http://index.okfn.org/place/ but http://index.okfn.org/place/2015, and (2) in January 2016 http://index.okfn.org/place/ will point to something else that what it pointed at in December 2015.

This may be useful and perfectly understandable for humans, but very confusing for machines. For example, if someone builds and aggregator/enricher for the 2015 data this year and wants to attribute the source, they will point to http://index.okfn.org/place/, but in January 2016 the attribution will be broken.

Lots of interesting discussions coming up!

Makx.


> -----Original Message-----
> From: Phil Archer [mailto:phila@w3.org]
> Sent: 1 May 2015 10:21
> To: Makx Dekkers; 'Public DWBP WG'
> Subject: Re: Thought for URI section
> 
> Thanks for this Makx,
> 
> I'm aware of the work you're doing in the DCAT-AP update which is clearly
> relevant to this group. I imagine other members of the WG will be as
> interested as I am in what it comes up with.
> 
> If it decides it wants to recommend that new terms to be added to DCAT, this
> is the WG that is chartered to do that. As you know, a lot of work is going into
> the quality and usage work which is where any such changes would sit.
> 
> Perhaps we can have a chat about this when we're in Krems in a few weeks'
> time.
> 
> Phil.
> 
> On 30/04/2015 18:29, Makx Dekkers wrote:
> > Phil,
> >
> > Interesting issue. I'd like to mention that in work that is being done around
> the DCAT application profile for data portals in Europe
> (https://joinup.ec.europa.eu/node/137964/), it has been noted that in DCAT,
> the W3C recommendation for describing datasets, there is currently no
> recommended way to describe time series.
> >
> > As far as I am aware, the approach you suggest cannot be expressed in
> DCAT as it is now. For the Application Profile, there is a discussion on how to
> model time series in DCAT, and it is scheduled that a draft for public review
> will be issued in the middle of May.
> >
> > Makx.
> >
> >
> >> -----Original Message-----
> >> From: Phil Archer [mailto:phila@w3.org]
> >> Sent: 30 April 2015 18:28
> >> To: Public DWBP WG
> >> Subject: Thought for URI section
> >>
> >> I'm using this as a jotter so I don't forget.
> >>
> >> I'm writing a doc and want to link to the 2014 Global Open Data Index.
> >> Well, the index is at http://index.okfn.org/place/ and that shows the
> >> 2014 results. If I want last year's I go to http://index.okfn.org/place/2013/.
> >> But I want to link persistently to the 2014 set *today* that happens
> >> to be the current one. Presumably when they do next year's index that
> >> will be at http://index.okfn.org/place/ and the 2014 data will be at
> >> http://index.okfn.org/place/2014 - but that doesn't exist yet.
> >>
> >> That's more evidence, IMO, that for any time series you always need a
> >> dated URI and a latest version URI.
> >>
> >> Rant over.
> >>
> >>
> >>
> >> --
> >>
> >>
> >> Phil Archer
> >> W3C Data Activity Lead
> >> http://www.w3.org/2013/data/
> >>
> >> http://philarcher.org
> >> +44 (0)7887 767755
> >> @philarcher1
> >
> >
> >
> 
> --
> 
> 
> Phil Archer
> W3C Data Activity Lead
> http://www.w3.org/2013/data/
> 
> http://philarcher.org
> +44 (0)7887 767755
> @philarcher1

Received on Friday, 1 May 2015 09:37:52 UTC