Re: [dxwg] Options for metadata to describe tabular data in a more structured way? (#1418)

A bit late to the game, but I agree with @kcoyle, @sabinem and several other that the main problem is knowing what `conformsTo` points to. Introducing subproperties like @makxdekkers are suggesting is a clear and simple way, but perhaps not flexible enough. I think it would be better to use a classifying property on the linked schema itself. The most obvious would be to use `rdf:type` and express the instance relation to a hierarchy of classes corresponding to a CSV on the web schema, XML Schema, JSON Schema, Datacube structure definition etc. However, introducing a hierachy of classes seems to be similar in rigidity to introducing subproperties. (From what I have seen profiles like DCAT-AP are not keen on introducing new classes, the preferred option seem to be to use vocabularies.) 

Hence, another alternative (which I also find a bit more appealing) would be to use two levels of `conformsTo`. E.g.

```
someDistribution                     dcterms:conformsTo  tabularDataSchemaInJSON
tabularDataSchemaInJSON   dcterms:conformsTo   CSVOnTheWebSpecification
```

Where CSVOnTheWebSpecification would be one of several well known data specification maintained as a vocabulary by those providing profiles of DCAT. The CSVOnTheWebSpecification could also be described as an instance of dcterms:Standard and perhaps utilize [the profiles vocabulary](https://www.w3.org/TR/dx-prof/) for a richer description.


-- 
GitHub Notification of comment by matthiaspalmer
Please view or discuss this issue at https://github.com/w3c/dxwg/issues/1418#issuecomment-971332240 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 17 November 2021 08:08:08 UTC