Various findings in examples

[rpietzsch]

cf. https://github.com/rpietzsch/data-product/pull/1/files

I tried to make sense of the examples and encountered some issues:

• converted to ttl for human use (if for technical purpose json is needed it is only one rapper / riot / jena call away)
  • the following problems can IMHO be better spotted in ttl
  • if you wish to include a json distribution of the examples I suggest to make ttl the authoritative format and generate the json on every push
• various non-existing concepts referenced, I'd put them all into dprod: although they do not seam to exist there yet either:
  • rdf:computedOn
  • rdf:isMeasurementOf
  • rdf:Metric
• changed rdf:title to rdfs:label (maybe dct:title is what you intended, but inconsistent with the other examples)
• avoid blank nodes for readability and because they are evil anyways ;-)

[nvar]

Thanks for the feedback!

1. Graph people talk turtle but most people prefer json that is why we left the examples as is.
2. DQV for computedOn, measurementOf etc. is an external open standard so the examples show how DQV can be used to describe quality on the data product. Similar with odrl. We are not including these standards in DPROD but we are showing how all these standards work together
3. Blank nodes are not my favorite as well and I always avoid them in implementation but for readability purposes we followed them in the examples

[rpietzsch]

• I agree to 1. but suggest to simply add both as - I guess - you are trying to cater graph and non-graph people
• reg. 2. I suggest to showcase an example integration with a real target ontology then; instead of making up non-existing concepts in defined namespaces
• reg. 3. fair enough

[jgeluk]

Is this issue still open?