Re: EXTERNAL: Re: Proposal for an additional term: mediaType

+Cc: Denny (Denny, we're discussing controlled values for
formats/carriers/mediatypes in schema.org markup, for learning and
bibliographic data)

On 21 September 2012 15:41, Suliman, Suraiya H
<suraiya.h.suliman@lmco.com> wrote:
> The list I have contains the following values. Note that this is not a complete list, just one from a particular publisher.

[snip]

Thanks for the list. I started going through looking these up in
Wikipedia, to get a feel for how close a match we might get.

Much of it is in there, but sometimes with a slight mutation of the
idea. I couldn't find 'Duplication Master' for example.

To give an idea,

http://en.wikipedia.org/wiki/Filmstrip
http://en.wikipedia.org/wiki/Flash_(photography)
http://en.wikipedia.org/wiki/Compact_Cassette
http://en.wikipedia.org/wiki/Calculator
http://en.wikipedia.org/wiki/Philips_CD-i
http://en.wikipedia.org/wiki/CD-ROM
http://en.wikipedia.org/wiki/Floppy_disk
http://en.wikipedia.org/wiki/Blu-ray_Disc
http://en.wikipedia.org/wiki/HD_DVD
http://en.wikipedia.org/wiki/DVD
http://en.wikipedia.org/wiki/Email
http://en.wikipedia.org/wiki/Slide_show#Digital_slide_shows (or just
http://en.wikipedia.org/wiki/Slide_show )
http://en.wikipedia.org/wiki/Field_trip ... and so on (I didn't try to
be complete, as your list was only itself indicative)

I've copied Denny from the Wikidata project, as we've talked before
about (i) having stable URL/URI references into the Wikipedia
universe, so that they can be used in data (ii) having simple type
hierarchy in Wikidata, so that CD-ROM, Floppy_disk, Blu-ray_Disk,
HD_DVD, DVD might be given a common Wikidata 'type' that we can cite.
Denny - any thoughts?

While any such finite list *could* be added directly into schema.org,
these are quite open-ended (anything that can carry information, plus
various types of thing related to the recording, creation or
collection of information). I don't think we could plausibly promise
to try to keep such an open-ended list up to date, and so maintaining
these value lists outside of Schema.org seems most prudent.

I know some publishers prefer to deal with simple (controlled) strings
/ labels rather than links (this came up in the final LRMI design
discussions) so I wouldn't suggest that publishers ought to write
"http://en.wikipedia.org/wiki/Philips_CD-i" where they currently just
have "CD-I". On the other hand, it does help to have less ambiguous
values, as well as the additional network of associations and metadata
that comes with Wikipedia (and Wikidata, Freebase, etc.).

The approach we suggested in
http://blog.schema.org/2012/05/schemaorg-markup-for-external-lists.html
is that in such cases, we should identify 'parent' types in
schema.org's hierarchy as attachment points for these external lists.
I think in this case we might also document an idiom for using strings
(preferred labels, in SKOS terms) from the external scheme. And then
we could point from schema.org to some external lists, ... maybe one
harvested from the LRMI community, the DC values, ONIX (e.g.
http://www.editeur.org/files/ONIX%20for%20books%20-%20code%20lists/ONIX_BookProduct_CodeLists_Issue_18_UTF8.txt
 ), etc. My instinct is that providers of such lists could choose to
use W3C SKOS + RDFa to publish their terms in machine-readable HTML.

Are there any more sources of raw values that we should look at here,
before we try to figure out how to group / link them?

Dan

Received on Friday, 21 September 2012 15:50:01 UTC