DCAT parsing issue on carriage returns

[Stijn Van Hoey]

After the adaptations on the IPT, we still have issues on the carriage returns of the description fields. The reason is the requirement to use the 'next paragraph' options whenever a new paragraph (basically any enter) is required.

This results in parsing issues with the dcat feed towards the open data Vlaanderen (and Belgium), resulting in not appearing in the listing.

The dcat feed output with the errors is provides issues in the data sets PINK (amphibia), Dung beetles and Alien macro-invertebrates.

We can check the well-functioning of the dcat feed with http://id.vlaanderen.be/vodap_validator
@DimEvil @milotictanja @peterdesmet @LienReyserhove
Whenever we add new datasets, make sure to do a quick check on the validator by pasting the URL of the INBO dcat-feed (http://data.inbo.be/ipt/dcat) into the validator.

[Stijn Van Hoey]

I adapted the metadata for the data sets with the paragraph-issue and we do not have any parsing issues anymore. The result of the vodap_validator is: http://id.vlaanderen.be/vodap_validatorresults/2017-10-23T11:14:21Z/

The parsing issues are solved, but we still have some other errors in the report:

1. Catalog - error - dct:issued should be a literal typed as date or dateTime.
2. Catalog - error - dct:modified should be a literal typed as date or dateTime.

See xsd:datetime for the required format.
It is a matter of explicitly define the value: "2012-09-17T17:01+02:00" ^^ xsd:dateTime or
"1999-08-16"^^xsd:date. In other words, adding this to the date field, see also https://www.infowebml.ws/rdf-owl/Literal.htm

and some warnings:

1. License Document - warning - dct:LicenseDocument does not exist.
2. Agent - warning - dct:type is a recommended property for Agent.

@peterdesmet we should probably tackle these on the IPT implementation level?

[Peter Desmet]

Yes, we should tackle these on an IPT implementation level. Maybe interesting to do as a 1 day project for a Java dev... or we could try ourselves.