Dear all,
I find it worrying that terms specifically selected/optimized for annotating web pages for SEO and SERP are recommended in a core vocabulary. A core vocabulary by its definition should be broadly applicable...
Understandably, the schema.org specification has a broad coverage, is made very accessible and is very well documented (to have as much as possible uptake, so search results get better and better).
Nevertheless, I would expect terms from 'specialized' vocabularies (at least not from a vocabulary that wants to cover everything). This would allow accurate descriptions of the concepts presented in the core vocabulary rather than cherrypicking terms from a Google search giants approved mishmash of other vocabularies when a good match could not (immediately) be found in other vocabularies.
So I disagree with the fact that for some concepts only schema recommendations are made and would argue to include (a table with) recommended terms for both general purposes and the annotation of web pages for each property (and class).
For example:
| Recommended RDF Term | General Purpose | Web Page Annotation (schema) |
| contactPoint | org:siteAddress | schema:contactPoint |
| logo | foaf:logo | schema:logo |
| ... |
Comments
Thanks for this, Laurens.
When schema.org first appeared I would have agreed with you, however, as it has evolved, I no longer do. At least, I don't agree that we should treat schema.org purely as an SEO artifact. Yes, it covers a lot of ground and is a general vocabulary. However, things like contact point *are* very general. To take the two examples you give.
It's true that org:siteAddress could be used (the spec makes it clear that the address need not be a physical address, however, schema:ContactPoint actyally suits us better than vCard as it includes hoursAvailable and productSupported that match the needs for CPOV (and other Core vocs) well.
foaf:logo or schema:logo is a straight forward choice. Given the popularity, nay, dominance, of schema.org, it seems sensible to go with the larger community.
schema.org is developed in an open environment (a W3C Community Group) with a documented process for adding terms to the snapshots. It has mass adoption and we do ourselves no harm by using it in my view. Finally, of course its primary purpose is to help search engines organise Web pages, but I imagine the vast majority of people will find public organisations by doing general Web searches, and therefore, where the needs of a Core Voc are met by schema.org, I would argue we benefit from using it.
This was discussed in the final WG meeting on 17/11 with the resolution that:
It is worth noting that as a direct result of a discussion around the CPOV's proposed use of schema:ContactPoint and opening Hours, it is schema.org that is being (very slightly) changed to align with the CPOV's needs. See https://github.com/schemaorg/schemaorg/issues/1444