Re: DQV, ISO 19115/19157 and GeoDCAT-AP - Issue-202 - representing conformance

Hi Andrea,

Thanks for the feedback! Comments after your email.

>>
>> Some time has passed since that last email.
>> I'd like to have a quick check before we close a number of actions
>> such as
>> https://www.w3.org/2013/dwbp/track/actions/262
>
> +1. Thanks!
>
>> Of notable interest for this discussion:
>>
>> - the latest version for the section on conformance
>> http://w3c.github.io/dwbp/vocab-dqg.html#ExpressConformanceWithStandard
>>
>> - updated diagrams showing the relations between metrics, annotations,
>> policies, etc.
>> http://w3c.github.io/dwbp/vocab-dqg.html#model
>
> Thanks again, Antoine. I checked the new specification, but I'm afraid I have to study it more thoroughly to have a good enough understanding of its use. The examples are very helpful, but of course they cannot cover all the use cases.
>
> So, I wonder whether you can help me understand one issue.
>
> I see (fig. 2) that the "circle" has been closed with prov:wasDerivedFrom. However, it is not completely clear to me how/whether it is possible to state that a resource *does* not conform to a given standard / quality policy. E.g., by using dqv:QualityAnnotation I can specify the tested resource (oa:hasTarget) and the result of a conformance test (oa:hasBody), but then I miss how I can state that the conformance test has been carried out against a given standard / quality policy.
>
> Overall, what is not completely clear to me is whether DQV has a kind of "expanded" representation of dct:conformsTo (on the lines of the Dublin Core to PROV mapping).


No we don't have such an expanded representation. More precisely, we defer to the one of GeoDCAT-AP (thus implicitly to the mapping between PROV-O and DC).
We just can't offer better. We don't have the resource to investigate this issue. Plus I'm really not sure it's worth it. Actually if we embark on this we would probably have to align with solutions for representing formal validation of data (a la RDF-Shapes) and this is both difficult and moving ground.

We could change the note at
http://w3c.github.io/dwbp/vocab-dqg.html#ExpressConformanceWithStandard
to be more precise about this. Here's a suggestion:
[
Finer-grained representation of conformance statements can be found in the literature, and applications with more complex requirements may implement them, including for example the requirement of representing 'non-conformance' tested by specific procedures. The GeoDCAT Application Profile, for example, suggests a "provisional mapping" for extended profiles, which re-uses the PROV data model for provenance (see Annex II.14 at [GeoDCAT-AP]). Such patterns come however at the cost of having to publish and exchange representations that are much more elaborate. They will also have to be aligned with the result of another ongoing efforts on data validation and the reporting thereof, as currently discussed around SHACL, for example. We have thus decided to postpone addressing these requirements for now.
]

I reckon this is not extremely satisfactory, but unless someone gives us a good solution to represent all this, I'm afraid there's no way to solve this in the coming weeks.

Cheers,

Antoine

Received on Thursday, 14 April 2016 03:37:48 UTC