- From: Richard Wallis <richard.wallis@dataliberate.com>
- Date: Tue, 29 Nov 2016 10:21:07 +0000
- To: Karen <karen.cravens@gmail.com>
- Cc: "schema.org Mailing List" <public-schemaorg@w3.org>
- Message-ID: <CAD47Kz70kbeH0NuB3pHgmpXtr5a7o3m7AG+eAxV0dtLdj1xu7Q@mail.gmail.com>
Hi Karen, I think you have identified a bit of a confused example. At first it looks like a reasonable description of a *MusicGroup*, with *event*s and *track*s. However as you have spotted, not only do some of the nested *MusicRecording* types validly have *interactionStatistic* properties, the *MusicGroup* also has one. Which is not valid. I also note that there is a *video* property which is not valid for *MusicGroup*. The question obviously is, what was the intension of creator of the example. Looking at the other serialisations, including the surrounding html, I believe the ‘interaction’ being described (for the *MusicGroup*, not the *MusicRecording*s) is probably better described using a *review* property, taking a *Review* type. As to *video*, that is slightly more complex. The least radical option would be to describe the *VideoObject* separately indicating that it is *about* the *MusicGroup* plus other properties indicating *creator*, *actor*, etc. More radical, but possibly worth pursuing would be the question as to if the domain of *video* should be expanded to include *Thing*, or alternatively the range of *image* be expanded to include *VideoObject*. It could be argued in these days of Snapchat, YouTube, etc. anything can have a *video* [image] as much as it can have an *image*. Whatever, the example (which appears 10 term pages on the Schema.org site) you identify needs fixing and I have raised an issue (#1446 <https://github.com/schemaorg/schemaorg/issues/1446>) to track that. ~Richard Richard Wallis Founder, Data Liberate http://dataliberate.com Linkedin: http://www.linkedin.com/in/richardwallis Twitter: @rjw On 28 November 2016 at 18:42, Karen <karen.cravens@gmail.com> wrote: > In the course of writing a Perl module to parse JSON-LD, I have been > feeding the examples to it as a test. Tried for an embarrassingly long time > to figure out why this MusicGroup entry was failing the tests before > realizing it's probably not supposed to be a MusicGroup. > > It appears on https://schema.org/interactionStatistic among other places > - and AFAICT MusicGroups never get interactionStatistics. Am I missing > something, or is that just a wrong @type? >
Received on Tuesday, 29 November 2016 10:21:43 UTC