- From: Greg Hullender <greg_hullender@hotmail.com>
- Date: Sat, 22 Oct 2016 23:31:33 +0000
- To: Richard Wallis <richard.wallis@dataliberate.com>
- CC: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>, Thad Guidry <thadguidry@gmail.com>, schema.org Mailing List <public-schemaorg@w3.org>
- Message-ID: <BL2PR20MB07710B657B291A4302DC9D88E2D70@BL2PR20MB0771.namprd20.prod.outlook.com>
Thanks! That looks well-thought-out. I have a feeling Google’s structured-data app won’t respect those, though. I run a web site that reviews short science fiction. We make extensive use of structured data in order to get Google to generate smart snippets. Here’s an example of a single review (one of thousands): http://www.rocketstackrank.com/2016/10/The-Eye-of-the-Swan-Kelly-Robson.html I’ve attached our document that describes how we used structured data, if anyone is interested. Google likes the markup we do; they display our smart snippets and our pages do well in searches, often outranking the authors’ own web pages. So in that sense, we don’t actually have a problem. Not right now, anyway. However, we like to do things right. If nothing else, it offers some protection against the future. Right at the moment, the things I worry about are: 1) We make each short story a CreativeWork because there is no more precise type for text. 2) We make anthologies PublicationIssue not Book because books must have authors. Also, if I recall correctly from the last time I tried this, Google rejects the idea that a book can have creative works as parts. 3) We don’t mark up our monthly summaries at all, even though they’re rich with data, because each entry is an abbreviated review of a short story, so none is quite complete as far as the Review type goes. http://www.rocketstackrank.com/2016/10/october-2016-ratings.html We’ve put a good bit of effort into trying to get all of this right, so it’s painful to have things that (to my way of thinking) can’t ever be right. Or which I simply don’t know how to make right. If there’s a better way to do this, I’d be very happy to hear it. Again, it’s not urgent; what we’re doing now works very well for us. We just don’t want it to abruptly stop working because Google make their system more sophisticated. --Greg Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986> for Windows 10 From: Richard Wallis<mailto:richard.wallis@dataliberate.com> Sent: Saturday, October 22, 2016 12:44 PM To: Greg Hullender<mailto:greg_hullender@hotmail.com> Cc: Stian Soiland-Reyes<mailto:soiland-reyes@cs.manchester.ac.uk>; Thad Guidry<mailto:thadguidry@gmail.com>; schema.org Mailing List<mailto:public-schemaorg@w3.org> Subject: Re: Leaping aliases or Property aliases in Schema.org On the issue of “ schema.org<http://schema.org/> doesn’t distinguish between the abstract creative work and the physical realization of it.” This is handled in the bibliographic world using FRBR<https://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records> (Functional Requirements for Bibliographic Works), using a hierarchy of Works, Expressions, Manifestations and Items. This was considered in depth in the Schema Bib Extend Group which introduced the exampleOfWork and workExample properties into schema:CreativeWork.<http://CreativeWork.> It is with this simple mechanism supplemented by MTEs it is possible to represent most any entity instance in that hierarchy. So “To Kill a Mockingbird”, the creative work independent of its production in any form would be a schema:CreativeWork with author, etc.. It would have a workExample relationship with “To Kill a Mockingbird”, a schema:Book with publisher, ISBN, numberOfPages, BookFormatType, etc. To describe the availability of an instance of such a schema:Book it would be multi-typed to be also a schema:Product enabling the capability to describe height, weight, image (cover image), and offers to purchase, loan, etc. An individual copy of “To Kill a Mockingbird” (the one signed by the author, or an individual copy on a library shelf) would be an MTE combination with schema:Book & schema:IndividualProduct which introduces serialNumber (barcode in library terms). There are examples of parts of this on the Book page: http://schema.org/Book#offeredBy-1 Also on that page is an example <http://schema.org/Book#bib-3> of how a series or multi-part work would be described, using hasPart and isPartOf to relate the volumes of “Lord of the Rings” together in the overall work. Note the use of a combination of schema:Book & schema:PublicationVolume. Referencing you anthology example, each story could be described as an individual schema:CreativeWork with author etc. and asPartOf the overall schema:CreativeWork (or subtype thereof) with publisher etc. ~Richard. Richard Wallis Founder, Data Liberate http://dataliberate.com Linkedin: http://www.linkedin.com/in/richardwallis Twitter: @rjw On 22 October 2016 at 17:07, Greg Hullender <greg_hullender@hotmail.com<mailto:greg_hullender@hotmail.com>> wrote: This is part of a larger problem: schema.org<http://schema.org> doesn’t distinguish between the abstract creative work and the physical realization of it. We’d like to say that the physical work comprises the cover (which has an artist) and the abstract text (which has an author). Similarly, the novel “To Kill a Mockingbird” has a length, an author, a copyright date, etc., but it doesn’t have a cover picture, a publisher, or an ISBN. A particular edition of “To Kill a Mockingbird” has a publisher, a publication date, a number of pages, and contents. The contents include the novel but also likely include an introduction by someone other than Harper Lee and might include study materials as well. It would be nice to separate the descriptions of the cover, the introduction, and the novel text from the description of the physical volume that combines them. The same problem occurs when trying to describe an anthology. It’s a book with no author but each story in it has an author. Magazines sort of work, via publication issue, but that’s a ad hoc solution to a more general problem. In general, I wish Schema.org made a distinction between a WrittenWork and a PublishedWork. WrittenWork would have a title, author, length in words (not pages), copyright date (not publication date), optionally might have editor, translator, translationOf, revisionOf PublishedWork would have an editor (but not an author), a length in pages, a publication date, a publisher, contents (creative works), and optionally might include illustrations, a volume number, etc. A WrittenWork might be part of a series. (E.g. “The Fellowship of the Ring” is part of “The Lord of the Rings.”) A PublishedWork might be part of a set. (E.g. Volume I and Volume II would be part of the same set.) It might be nice to have a “publishedAs” field in WrittenWork to allow enumeration of places the work was published. --Greg Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986> for Windows 10 From: Stian Soiland-Reyes<mailto:soiland-reyes@cs.manchester.ac.uk> Sent: Saturday, October 22, 2016 5:48 AM To: Thad Guidry<mailto:thadguidry@gmail.com> Cc: schema.org Mailing List<mailto:public-schemaorg@w3.org> Subject: Re: Leaping aliases or Property aliases in Schema.org Perhaps you could amend with some owl:equivalentProperty rule. (Assuming you mean the "artist" of a CoverArt must be a "cover artist".) But would not the cover artist property go from the Work (e.g. a book or music album) rather than from the CoverArt itself? On 20 Oct 2016 2:14 am, "Thad Guidry" <thadguidry@gmail.com<mailto:thadguidry@gmail.com>> wrote: So...There's this... https://bib.schema.org/CoverArt that has an 'artist' property. That 'artist' property in actuality could be a 'cover artist' that could be mapped to https://www.wikidata.org/wiki/Property:P736 But I have no reliable way to do that because we don't have Aliases at a Type/Property level. Bummer. How do we fix or solve this ?
Attachments
- application/pdf attachment: Rocket Stack Rank Structured Data.pdf
Received on Saturday, 22 October 2016 23:32:06 UTC