Closed pvgenuchten closed 3 years ago
Thanks for reporting this issue, @pvgenuchten .
I guess this is the pattern/template defined in the Swedish profile of DCAT-AP (I don't know if other DCAT-AP profiles recommend the same one).
If this is the case, this issue should be preferably addressed by a custom version of the GeoDCAT-AP XSLT, implementing the requirements of the Swedish profile.
/cc @matthiaspalmer
Ok, clear answer
The DCAT-AP specifications (both earlier and latest) are rather vague regarding how to express the vcard:Kind of the dcat:contactPoint. In the Swedish profile we thought it better to provide guidance for implementors on which properties to use, e.g. which properties to use and weather to use the n-ary or direct properties.
Clearly, when you implement a mapping you have to decide, which is apparent in the transform mentioned XSLT file. And when you build something like a portal you typically need to either harmonize or understand a lot of different cases. And apparently we made different decisions in the Swedish DCAT-AP compared to the Inspire to dcat XSLT.
I just checked the German and Norwegian DCAT-AP profiles, they do not seem to provide any guidance on this unfortunately. I guess a deeper look is required to see if there is a majority rule to gravitate towards in the future.
Swedish validator fails at
https://github.com/SEMICeu/iso-19139-to-dcat-ap/blob/e7aa2d802f6e337ffce64c9167c66b835916f421/iso-19139-to-dcat-ap.xsl#L1507
expects hasTelephone as https://www.w3.org/TR/vcard-rdf/#Examples