W3C home > Mailing lists > Public > public-vocabs@w3.org > October 2013

Re: CreativeWork relationships

From: Wallis,Richard <Richard.Wallis@oclc.org>
Date: Mon, 14 Oct 2013 14:29:55 +0000
To: Antoine Isaac <aisaac@few.vu.nl>
CC: Charles McCathie Nevile <chaals@yandex-team.ru>, "<public-vocabs@w3.org>" <public-vocabs@w3.org>, "Wallis,Richard" <Richard.Wallis@oclc.org>, "public-schemabibex@w3.org" <public-schemabibex@w3.org>
Message-ID: <1A4454A4-3280-4783-AC2F-0BF6698BD011@oclc.org>
As Chaarls points out there are a few types of these relationships being grouped together here, some more concrete than others.

There are what I would term the known derivation relationships between instances of the same thing (this is an example of that, the hardback book edition of the work, etc.)
eg.
<http://www.worldcat.org/oclc/57714007>
a schema:Book;
schema:name "Harry Potter and the half-blood prince"@en;
schema:author <http://viaf.org/viaf/116796842>;
schema:exampleOf <http://en.wikipedia.org/wiki/Harry_Potter_and_the_Half-Blood_Prince>
.

There are the less exact relationships - isBasedOn, commonEndeavour, etc.
eg.
<http://www.imdb.com/title/tt0417741/<http://www.worldcat.org/oclc/57714007>>
a schema:Movie;
schema:name "Harry Potter and the half-blood prince (2009)"@en;
schema:director <http://www.imdb.com/name/nm0946734<http://viaf.org/viaf/116796842>>;
schema:isBasedOn <http://en.wikipedia.org/wiki/Harry_Potter_and_the_Half-Blood_Prince>;
schema:commonEndeavor <http://www.worldcat.org/oclc/57714007>
.

There are derived relationships - translationOf, representationOf, etc.
eg.
<http://en.wikipedia.org/wiki/Venus_de_Milo>
a schema:Sculpture;
contentLocation "Louvre Museum, Paris, France";
schema:dateCreated "Approx. 130-100 BC"
.
<http://www.flickr.com/photos/elhombredepan/4708103537/>
a schema:ImageObject;
schema:creator <http://www.flickr.com/photos/elhombredepan/>;
schema:isRepresentationOf <http://en.wikipedia.org/wiki/Venus_de_Milo>;
schema:dateCreated "2010-04-07"
.
I've added isRepresentationOf to the list.


Taking a stab answering Chaarls specific examples:

West Side Story, the Musical, is a version of Romeo and Juliet. Is this what you mean by "isBasedOn"?

It is either isBasedOn or isAdaptionOf


West Side Story, the movie, is a version of the Musical, but there is a change to make the format. Is this a format of West Side Story, and is using a

It would be based on, or an adaption of, the musical

Wikipedia entry URL as the source a sensible thing to do?

Depends on the Wikipedia description I suppose - is it the generic work it is describing or a specific example (the musical or the film)


Blade Runner is an adaptation of Do Androids Dream of Electric Sheep, and often called the book of the movie. But it is not really as close to the original as West Side Story is to Romeo and Juliet. Film adaptations are often pretty different from their original sources. Shirley Valentine is another example of a movie and play that are not all that close to each other, despite being very recongisably the same work.

May be we need 'wasInspiredBy'


Romeo and Juliet, the Baz Luhrmann movie, is a version of Romeo and Juliet in a different format. Is this a workExample?

I do not know the Baz Luhrmann movie, but if the text spoken by the actors is that of the play, yes I think it would be.  In the same way that an audiobook would be a workExample of the book.  If however the text had been adapted for the movie (making it a different work, with an additional author) then it would be basedOn.


There are also various text versions of Romeo and Juliet. I'm guessing they are all candidates for workExample/exampleOfWork.

In principle yes. These well known, often copied, often historical works are a challenge for such analysis as they are often changed in the copying and translation enough to make them not the original work.  With the fairly concrete wokExample and the more vague adaptionOf and basedOn she should be able to satisfy the needs of most cases.  Examples & documentation should help here.


The other set of use cases are where there is no clear "original".

For example, I write a new, interactive internet site telling a story about swans and ostriches. It is natively video, or text, or text and audio, and so on - you decide what pieces you actually view.

I could describe the resource as having multiple versions, each with different characteristics (format, language, ...) but the URL is the same. I think I want to use named graphs to distinguish these cases. (I want to do the same thing to handle the fact that there are effectively multiple versions of the resource with different accessibility characteristics, but the URL where you get the resource is the same).

Depends on how stand-alone each component is.  I would suggest each part is an individual CreativeWork, with probably a commonEndeavor relationship, and they are all partOf a group work.

I can't see us getting named graph capability in Schema anytime soon ;-)


I think it would be useful to flesh out my use cases a bit more, so they can be worked up into concrete examples with answers. But so far I don't see:

- How do I handle the distinction between "X is an original. Y is based somehow on X" and "X has multiple representations because it is true multimedia"?
- What is the difference between a play or film that is an adaptation of a book, and something that is just a retelling of the same story? How does that compare with a resource that is an attempt to adapt as literally as possible in a different format? (And how does that compare with translation - is the best translation of the Odyssey alliterative verse or blank verse or hexameter or prose)?

Much of this is down to the metadata creator - we need to give them enough tools to do the job but not too many to cause significant confusion.


~Richard

On 14 Oct 2013, at 09:11, Antoine Isaac <aisaac@few.vu.nl<mailto:aisaac@few.vu.nl>> wrote:

Hi all,

For what it's worth, here are examples of 'core' properties that can hold between Creative work in our data model in the Europeana initiative:
-- dc:subject and edm:isRepresentationOf for "aboutness" links (Mona Lisa and a historical picture of Mona Lisa that represents it)
-- edm:realizes, which is quite FRBR-related (An item of the Gutenberg’s edition realizes the Bible)
-- edm:isSimilarTo: covering true and "questionable" cases of derivation in the FRBR sense. Including sub-properties edm:isDerivativeOf (for real derivation cases like re-working, extension--this is a sub-property of dcterms:isVersionOf, targeted perhaps rather at CreativeWorks) , edm:incorporated (for inclusion / re-use of works within others) and edm:isSuccessorOf (for "sequels")

I can contribute more documentation and example for these, but the mail would be long then. So will wait till there's interest. For the curious, however, you can have a look at our quite dry PDF reference at [1].

Then you can have more general links (dc:relation), general part-whole relation (dcterms:hasPart), citation (dcterms:references). These ones are clearly in schema.org<http://schema.org> already.

@Chaals: the criteria for judging what counts what as an intellectual work, and thus a derivation, a version, or a mere change of format, are very fuzzy. I'm not sure we can give one answer. E.g. for translators of historic texts a translation would surely count as a creative work, and what is a good translation is a matter of research-level debates. For more recent book publishing cases, the translation may not be such a big deal... but not always: in French the first translation of Edgar Poe was done by Charles Baudelaire, and it is acknowledged an important work for the later fame of both authors.

Things may be more straightforward for format-specific variants, as you write: "resource that is an attempt to adapt as literally as possible in a different format", as in accessibility scenarios.
For this a property like dcterms:hasFormat could be used. And I think the different variations should have their different URIs. Maybe one want to serve a common access URI for them (say, typed as an OAI-ORE Aggregation [2]) but that's a story slightly different than creating the RDF graph for these different versions, their attributes and the links between them, which needs separate URIs.

Cheers,

Antoine

[1] http://pro.europeana.eu/documents/900548/0d0f6ec3-1905-4c4f-96c8-1d817c03123c
[2] http://www.openarchives.org/ore/primer

Hi Richard,

On Mon, 14 Oct 2013 00:15:43 +0200, Wallis,Richard <Richard.Wallis@oclc.org<mailto:Richard.Wallis@oclc.org>> wrote:

One of the topics we have discussed in the SchemaBibEx group has been properties to describe some relationships between CreativeWorks.

In several recent threads on public-vocabs there has also been mention of relationships between CreativeWork such as adaptionOf, isBasedOn, etc.

It is clear therefore that this topic is more broadly relevant than the bibliographic focus of SchemaBibEx

Yep.

To help the discussion in this area I have produced a small list of these proposed properties for CreativeWork, which have CreativeWork as a range.

As it was assembled by trawling threads which I remembered included such mentions, it is almost certainly not a comprehensive list, it also includes some that are very similar in intention.

Yes. In particular, I have two different sets of use cases in mind. One revolves around understanding that something is like something else, but really not the same.

West Side Story, the Musical, is a version of Romeo and Juliet. Is this what you mean by "isBasedOn"?

West Side Story, the movie, is a version of the Musical, but there is a change to make the format. Is this a format of West Side Story, and is using a Wikipedia entry URL as the source a sensible thing to do?

Blade Runner is an adaptation of Do Androids Dream of Electric Sheep, and often called the book of the movie. But it is not really as close to the original as West Side Story is to Romeo and Juliet. Film adaptations are often pretty different from their original sources. Shirley Valentine is another example of a movie and play that are not all that close to each other, despite being very recongisably the same work.

Romeo and Juliet, the Baz Luhrmann movie, is a version of Romeo and Juliet in a different format. Is this a workExample?

There are also various text versions of Romeo and Juliet. I'm guessing they are all candidates for workExample/exampleOfWork.

The other set of use cases are where there is no clear "original".

For example, I write a new, interactive internet site telling a story about swans and ostriches. It is natively video, or text, or text and audio, and so on - you decide what pieces you actually view.

I could describe the resource as having multiple versions, each with different characteristics (format, language, ...) but the URL is the same. I think I want to use named graphs to distinguish these cases. (I want to do the same thing to handle the fact that there are effectively multiple versions of the resource with different accessibility characteristics, but the URL where you get the resource is the same).

I think it would be useful to flesh out my use cases a bit more, so they can be worked up into concrete examples with answers. But so far I don't see:

- How do I handle the distinction between "X is an original. Y is based somehow on X" and "X has multiple representations because it is true multimedia"?
- What is the difference between a play or film that is an adaptation of a book, and something that is just a retelling of the same story? How does that compare with a resource that is an attempt to adapt as literally as possible in a different format? (And how does that compare with translation - is the best translation of the Odyssey alliterative verse or blank verse or hexameter or prose)?

cheers

Chaals

Comments on the list, omissions, suggestions are invited so that we can get some consensus around a hopefully small set, to cover many circumstances, to propose as additions to CreativeWork.

The list is currently hosted on the SchemaBibEx Wiki: http://www.w3.org/community/schemabibex/wiki/Schema_CreativeWork_Relationships

~Richard.
Received on Monday, 14 October 2013 14:31:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:32 UTC