Failed aggregateRating validation of Google's example by SDTT

[e-orlov]

Try to validate with SDTT the first example from this url: https://developers.google.com/structured-data/rich-snippets/reviews, this one:
`

Super Book

Ultra interesting. Super impressive.

Book rating: 88 out of 100 with 20 ratings

`

You'll get an error The property aggregateRating is not recognized by Google for an object of type Thing.
It seem to be a Schema.org issue - here, https://schema.org/AggregateRating, is written, that Thing since unknown time doesn't allowed to be rated/reviewed, but only Place, Event, Service, CreativeWork, Product, Brand, Offer, Organization

[vholland]

What is the Thing you are describing? I am trying to think of a case where it has reviews, but is not some type of Place, Event, Service, CreativeWork, Product, Brand, Offer, or Organization.

[DDeering]

@e-orlov The issue is actually not on schema.org's side. Google's Structured Data Testing Tool (Google's tool, not schema.org's) just does not accept that particular markup for a Thing. Change Thing to Product, Organization, Service, etc and it tests fine. The markup itself is fine. Google's testing tool just has not been set up to accept it.

[danbri]

While the specifics here (posting examples which don't valid with a checker on the same site) are a Google issue, there is still an underlying confusion coming from schema.org:

• http://schema.org/itemReviewed (from AggregateReview or Rating to Thing), vs
• http://schema.org/aggregateRating (from Brand, CreativeWork, Event, Offer, Organization, Place, Product, Service, to an AggregateReview)

Also, any http://schema.org/Rating is somehow tied to a Review, via http://schema.org/reviewRating.

It seems to me that (1.) it should be possible to review/rate more or less anything, and as @vholland suggests the list Brand, CreativeWork, Event, Offer, Organization, Place, Product, Service comes close to approximating "Thing". This list omits Intangible, Person, Action and MedicalEntity, of which Person seems the most interesting and controversial. I don't think we are in a hurry to encourage schema.org use for rating people. (2.) the direction in which a relationship is named is logically arbitrary, but in practice we all tend to let it influence our thinking: there is a waryness of 'cluttering' Thing with out-going properties, while being more liberal in tolerating properties whose values are Things.

[danbri]

Updated title from "Failed aggregateRating validation of Google's example by SDTT" to capture the most obvious work item for schema.org (rather than Google).

[danbri]

Hmm, I said that I changed the title but it didn't seem to stick. I don't see an obvious schema.org issue here, but I believe the general issue around ratings/reviews and the list of things those can be attached to is already discussed in several other issues, notably #637 so I will close out this issue in hope of concentrating attention in a smaller list of places. Thanks @e-orlov