W3C home > Mailing lists > Public > public-vocabs@w3.org > March 2015

Re: Proposal for Schema.org extension mechanism

From: Guha <guha@google.com>
Date: Wed, 4 Mar 2015 16:12:21 -0800
Message-ID: <CAPAGhv-U7QOuMsFk=WMRAiHkTMPvLmTmeeJ-URK68b4G6Fd8=w@mail.gmail.com>
To: Wes Turner <wes.turner@gmail.com>
Cc: Mark Baker <distobj@acm.org>, James M Snell <jasnell@gmail.com>, Ralph Swick <swick@w3.org>, Sandro Hawke <sandro@w3.org>, Tim Berners-Lee <timbl@w3.org>, W3C Vocabularies <public-vocabs@w3.org>
First an apology. schema.org/course is a very badly named relation, that
needs to be renamed/deprecated. And the person who let that happen (me)
should be severely reprimanded.

To answer the more general question, extensions should not reuse terms to
mean something completely different. All of core schema.org is in every
extension, in that extensions 'namespace'. Since we would not want two
completely different meanings of schema-ext123:course, we should have
schema-ext123:course mean something completely different from
schema.org/course. They are the same!


On Wed, Mar 4, 2015 at 3:37 PM, Wes Turner <wes.turner@gmail.com> wrote:

> Here's a practical question:
> * schema.org/Course is in development for schema.org Core [1]. For
> purposes of discussion, say that it wasn't; that schema:Course is part of
> an in-development extension
> * schema.org/course already has a defined domain and range (schema:Place)
> that is somewhat subjectively disjunctive from schema:Course, for online
> courses (e.g. from edX)
> How is this resolved with the proposed extension mechanism?
> Is there a schema-ext123:course which could have a range of schema:Course?
> [1] https://github.com/schemaorg/schemaorg/issues/195
> On Wed, Mar 4, 2015 at 1:42 PM, Guha <guha@google.com> wrote:
>> Very good point. I should have been more clear.
>> Use of schema.org terms has evolved from only search to many other
>> applications, both from schema.org sponsors (e.g., Google Now, Cortana)
>> to non-search companies (Rich Pins from Pinterest). These applications tend
>> to have narrower domains than search and require more specialized
>> vocabularies. Many of these more specialized vocabularies, though more
>> specialized, also tend to include many of the general terms found in
>> schema.org.
>> We are simply trying to create a mechanism for the to reuse these terms
>> by extending schema.org. Of course, they could do their own thing and
>> rely on webmasters figuring out namespaces, which terms from which
>> namespace, etc. But experience has shown that webmasters often find that
>> too onerous. This is an attempt at a solution for that.
>> And of course, as always, use of vanilla RDFa is always welcome.
>> Hope this clarifies.
>> guha
>> On Wed, Mar 4, 2015 at 11:35 AM, Mark Baker <distobj@acm.org> wrote:
>>> Hi Guha, everyone...
>>> On Sun, Feb 22, 2015 at 8:24 PM, Guha <guha@google.com> wrote:
>>> > For
>>> > something to go into the core, it not only has to meet the quality
>>> bar, but
>>> > also be of general interest to a substantial fraction of the internet
>>> > community. This second condition does not hold for reviewed extensions.
>>> This is a very interesting statement to me. It's certainly possible
>>> that I'm missing the point here, as although I've been interested and
>>> supportive of (and written code to support) the schema.org effort,
>>> I've only recently joined this list. But it's been my understanding
>>> since it was first announced that the value here is basically a
>>> (potentially temporary) trade-off of decentralized evolvability for
>>> ease/speed of deployment. So the search engines basically say via
>>> schema.org, "Here's some general terms that we understand", and
>>> publishers down-convert their content from vertical-specific terms to
>>> those more general terms.
>>> The motivation in the current proposal states, "As schema.org adoption
>>> has grown, a number groups with more specialized vocabularies have
>>> expressed interest in extending schema.org with their terms". I don't
>>> doubt that's true, but as we all know, the driving force that makes
>>> the schema.org proposition valuable isn't from "groups", it's the
>>> search engines and their support of those general terms. As the terms
>>> become more and more specific, of interest to smaller and smaller
>>> communities, and therefore of less interest to search engines to
>>> support, that value evaporates AFAICT.
>>> So I really wonder what the benefit is here. For those communities
>>> with vocabularies of a less general nature, why not just publish with
>>> vanilla RDFa? What value is there to be hitched to schema.org in the
>>> way described in this document?
>>> Thanks.
>>> Mark.
Received on Thursday, 5 March 2015 00:12:49 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 5 March 2015 00:12:50 UTC