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

Re: Redefine and reuse?

From: Dan Scott <dan@coffeecode.net>
Date: Wed, 24 Jul 2013 08:56:00 -0400
To: Martin Hepp <martin.hepp@ebusiness-unibw.org>
Cc: Owen Stephens <owen@ostephens.com>, kcoyle@kcoyle.net, public-vocabs@w3.org
Message-ID: <20130724125551.GA2261@denials.eastlink.ca>
On Wed, Jul 24, 2013 at 09:59:50AM +0200, Martin Hepp wrote:
> On Wed, Jul 24, 2013, Owen Stephens wrote:
>> Hi Karen,
>> How would you see serialNumber mentioned by Wes mapping against library
>> item identifiers?

> The serialNumber is in theory meant as a unique item identifier
> assigned at production or market entrance time, like the serial number
> of a digital camera, the Vehicle Identification Number of a car, etc.
> A library-specific identifier (like shelf locator ID) is not really a
> serial number. I do not see major problems with using serialNumber
> this way, but it is not exactly as initially meant.

(which initiated Karen's question to this list), I had suggested mapping
serialNumber to the item barcode (which is a unique item identifier, at
least in a library context for physical items), sku to the call number,
gtin13 to the ISBN13 (where available), and more.

serialNumber probably wouldn't be an applicable property for a new book

I feel like some of the specifics of the Schema Bib Extend group
proposals are spilling out a bit prematurely on public-vocabs... while
early feedback is good, if you're interested in helping us hash out the
specifics please join us at http://www.w3.org/community/schemabibex/ so
that we can come back to public-vocabs with strong proposals!

Getting back to broader schema.org questions that might be more suitable
for public-vocabs at this time, Karen's basic "redefine" question was a
policy one: "Are the definitions of existing properties in schema.org
rigid, or are they open to being redefined to accommodate more use
cases?" I believe the answer we arrived at is "open to being redefined
to be more inclusive" with a caution against trying to make things so
generic as to be meaningless.

And the second question, on "reuse", was (I think) a technical one,
based on finding that the "/" extension mechanism has fallen out of
favour: "Is it possible to reuse an existing schema.org type for a new
type, while providing more natural domain-specific property names, in a
way that schema.org processors could handle gracefully?" A specific
example would be possibly mapping Offer.sku / Offer.serialNumber to,
say, Holdings.callnumber / Holdings.barcode.

And I think the emerging answer to this second question is based on
practical notions: rather than having to maintain what are essentially
copies of the same type/property vocabulary with different names for
various corners of the universe, reuse the existing type/properties
where the mappings are logical but provide best practices documentation
focused on the domain of use.

I may be over-simplifying this second answer, prematurely, and I
certainly have a bias towards documentation as a former technical
writer, but I think these two questions are important to anyone coming
to schema.org with a proposal. Possible fodder for the schema.org FAQ
and/or an updated "Extension mechanism" document?

There may be a slightly different discussion about defining your own
vocabulary and using RDFa Lite to mix in properties here, too, for those
groups who want to put their proposals into practice immediately, and a
corresponding discussion about namespaces.
Received on Wednesday, 24 July 2013 12:56:31 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:49:00 UTC