Re: BIBFRAME and schema.org

The Freebase schema is worth reviewing for those who aren't already
familiar with it.  It attempts to strike a balance between information
modeling purity and real world practicality.

On Fri, Jun 28, 2013 at 11:52 AM, Dan Scott <denials@gmail.com> wrote:

>
> During the last call, I proposed (via chat) that Freebase's
> adaptedWork / adaptedFrom properties might make more sense than the
> proposed hasInstance / isInstanceOf for expressing relationships
> between CreativeWorks. I'm not sure we really need a Platonic ideal /
> FRBR Work in schema.org; it seems to be a potential rat hole that
> would be better avoided, as the abstract "Work" is subject to
> revisionism and argumentation for little benefit to the linked data
> effort.
>
> For example: would the abstract CreativeWork for "The Little Mermaid"
> be the Disney creation? Surely not; it would be the Hans Christian
> Andersen work on which the Disney story was based, but it would not be
> the English translation; it would be "Den lille havfrue" - but wait,
> Andersen's work doesn't even include "Ariel" as a character's name,
> and surely the vast majority of people looking for "The Little
> Mermaid" actually want the Disney films / books / tv series / video
> games / figurines / stickers / whatever... and perhaps at some point
> in the future we will discover that Andersen's work was based on a
> previously existing oral tale. Do we even want to try to have to
> express that, and maintain that, when it seems much better suited to
> the realm of historical literature academics & their research papers &
> books & conference proceedings?
>
> In short, I don't think an abstract CreativeWork and all of the FRBR
> Work baggage that would carry offers significant benefits to our
> efforts. I, for one, would be happy to link off to, say, the wikipedia
> page on "The Little Mermaid" (either
> http://en.wikipedia.org/wiki/The_Little_Mermaid or
> http://en.wikipedia.org/wiki/The_Little_Mermaid_(1989_film) or
> http://en.wikipedia.org/wiki/The_Little_Mermaid_(franchise) or
> http://en.wikipedia.org/wiki/The_Little_Mermaid_(disambiguation)) or
> their Freebase equivalents, and let the linked data lead interested
> parties to explore the connections and arguments further.
>

Here's the Freebase equivalent of the first (the work):
http://www.freebase.com/m/01rx7z
with names in 27 languages:
http://www.freebase.com/m/01rx7z?i18n=<https://www.freebase.com/m/01rx7z?i18n=>
and links to 25 Wikipedias
links to 4 adaptations (Disney, anime, another animated film, & a stage
production): https://www.freebase.com/m/01rx7z#/media_common/adapted_work
14 book editions (which are, in turn, linked to OpenLibrary, Google Books,
etc): https://www.freebase.com/m/01rx7z#/book/book

The Freebase schema treats editions, adaptations, and translations as
separate kinds of links and this seems more natural to me than smushing
them altogether.  A book edition and a film adaptation are very different
beasts.  Adaptions can be to/from any media and can be anything from
remakes to works only loosely based on an original. Foreign language book
editions are linked both to the abstract work and to a translation node
that contains information about the translator, target language (probably
need a picture here), although this isn't well populated in the current
data because the information isn't typically available in machine readable
form.

The author's page: http://www.freebase.com/m/03j90 in addition to being
linked to 41 Wikpedia's is linked to the LC NAF, OpenLibrary, NNDB, IMDB,
NY Times, MusicBrainz (for spoken word recordings), VIAF, and a bunch of
book sites.

As Freebase, Wikidata, DBpedia, etc make progress, this tapestry of
connections will get richer and richer.  I'd argue that what people want is
a way to access and connect to this tapestry, not a separate "library" view
of the world.

Tom