Re: [dxwg] Usage notes [RUN] (#86)

Thanks for elaborating this proposal, @riccardoAlbertoni .

I would not opt for SKOS, as none of the documentation properties seems to denote "usage".

I would nonetheless support the use of a simple property, although not excluding the possibility of using a more articulated representation (in case people need it). So, I welcome the idea of providing a mutual mapping. BTW, as I mentioned during the last DCAT call, this links back to the issue around the mapping of qualified and non-qualified relationships (https://github.com/w3c/dxwg/issues/80).

For the simple property, an option ([documented in UC9](https://www.w3.org/TR/dcat-ucr/#ID9)) is `vann:usageNote`. 

Although VANN is meant to provide a way to annotate vocabularies, the definition of [`vann:usageNote`](http://vocab.org/vann/#usageNote) does not constrain its use to vocabularies. Quoting:

> A reference to a resource that provides information on how this resource is to be used.

The possible issue relates to a discussion  we had on the use of this property for annotating DCAT, around the question on whether this is a datatype or object property, since it is a subproperty of `rdfs:seeAlso`, which is supposed to be used as an object property (see https://github.com/w3c/dxwg/issues/233). As far as I know, `vann:usageNote` is being used in practice either with a "literal" (narrative text) or to point to a document. So, I'm personally not concerned to using it in DCAT for both options (as it could be done with SKOS documentation properties). I would refer to this option as **Solution 0**.

However, if the group decides that this option is not acceptable, I would be in favour of Solution 2 - namely, defining a property `dcat:usageNote`, that could be used either as a datatype or an object property. For this purpose, it can also be made a subproperty of `skos:note`.

About the alternative, more complex representation, I have yet to make up my mind on your 2 proposals, @riccardoAlbertoni - namely, using either OA (more generic) or DQV (focussing on data quality aspects). It is indeed true that usually this is about fitness for purpose, but there are also other cases, and I'm not sure all of them may fit into the notion of data quality (even considered in its broadest sense). So, maybe it is "safer" to use OA. But, again, at the moment I don't have a strong position in favour of / against either options.

Said that, I think the complex option needs to take into account another aspect. Both OA and DQV have been designed to enable the creation of annotations by anyone - providers and consumers. Since the use case/requirement we are discussing here is about giving data providers a way to specify usage notes, the annotation creator must be explicitly indicated.

Based on this, I would suggest we first take a decision on whether we would like to support or not the solution based on a simple property, and consider the more articulated option afterwards.

About the simple property, my vote is as follows (in cascade):
1. **Solution 0** (`vann:usageNote`)
2. **Solution 2** (`dcat:usageNote` with open range) if the group rules out Solution 0.

BTW, I've also contributed my vote accordingly, but it's incomplete, as Solution 0 is not included in the options.

-- 
GitHub Notification of comment by andrea-perego
Please view or discuss this issue at https://github.com/w3c/dxwg/issues/86#issuecomment-471114457 using your GitHub account

Received on Friday, 8 March 2019 23:37:03 UTC