Re: Normative reference to schema.org in EPUB Accessibility?

Hi folks

On Fri, 10 Sep 2021 at 11:03, Ivan Herman <ivan@w3.org> wrote:

> Ralph, Philippe,
>
> here is what we are thinking of doing:
>
> - a separate W3C community group will be set up, whose sole purpose in
> life will be to be the guardian of the schema.org a11y terms. The CG
> would take over (and clean up) [1]. Once the setup is done, we will also
> have to contact schema.org to make the situation clear (and put some sort
> of redirect from [1] to the new CG's site).
> - the a11y specification would directly, and normatively, refer to the
> schema.org vocabulary possibly referring to the Community Group's site,
> too. We believe that type of stability is important for the community.
> There is already a PR showing the differences in the spec[2]
>
> Is this o.k. with you?
>

This would be analogous to other areas where Schema.org often pulls in
suggestions from a dedicated community with its own (public, open to all
etc.) discussions.

Often but not always a W3C CG. For fact-checking schemas (Politifact etc.)
we (ie Schema.org) engage with experts via the international fact checking
network (ifcn). For education/learning, the LRMI project is now a part of
the Dublin Core Metadata Initiative (DCMI), etc. In each case from
Schema.org’s perspective it is great to have loosely-coupled collaborations
like this but we are generally conservative about making additive (rather
than usability/integration) edits unless they’re in the context of an
application that will actually use/consume the data. See
https://schema.org/docs/howwework.html and nearby.

So long as the CGs understand that they’re making proposals to an
independent project (Schema.org) and don’t have exclusive say over schema
designs, these structures can work well. We have found over the years that
all topics and domains are rather intermingled and that the idea of
definitively delegating areas is fraught with difficulty. For example -
medical schemas vs healthcare information vs factchecking vs local business
information vs life sciences; we have schemas in all these areas which have
touched on the coronavirus situation.

I should also mention that there are other areas of Schema.org that touch
upon Accessibility, beyond the handful of properties discussed here (eg
SpeakableSpecification, or anything wrt MediaObject). My advice would be
for the CG to have a scope that helps its participants engage on Schemas
for Accessibility in general, and that has equal emphasis on collaborations
to *consume* the data, since that’s the path to this data being useful

Cheers

Dan


> Thanks
>
> Ivan
>
>
> P.S. A an aside, the new CG would need a github repository under the w3c
> organization. I hope I have the green light to set up when the time comes.
>
>
> [1] https://www.w3.org/wiki/WebSchemas/Accessibility
> [2] https://github.com/w3c/epub-specs/pull/1808
>
>
>
> On 8 Sep 2021, at 22:05, Ralph Swick <swick@w3.org> wrote:
>
>
>
> On 2021-09-08 09:37 AM, Ivan Herman wrote:
>
> Ralph, Philippe,
> this type of question comes up regularly, but I did not see any clear cut
> answer.
>
>
> There's no absolute determination in advance; this is intentional.  Each
> case has its own considerations.
>
> The EPUB Accessibility spec[1] has a section on package metadata[2] to
> refer to metadata like access mode or accessibility features. The
> specification defines these terms in general, meaning that it is not
> properly defined which terms are to be used in a real metadata
> instantiation; this is left to the separate WG Note on a11y techniques[3]
> which reveals the thinly veiled fact that, in practice,
>
>
> "thinly veiled" is a big flag for me.  The spec should be clear and as
> precise as possible about the Working Group's intentions.  If the WG
> intends that the conformance expectations for an eventual W3C
> Recommendation maximize interoperability with specific metadata usage it
> should state so.  If it believes that the schema.org terms and their
> definitions are the correct solution, it should state so -- and be prepared
> to argue its position with the Director, the W3C Members, and the Community.
>
> these general terms refer to their equivalents in schema.org <
> http://schema.org>[4]. Indeed, all the terms defined in [2] are,
> actually, defined in schema.org <http://schema.org>, and those are the
> only mappings for those terms. Those terms are not out of the blue,
> actually: they have been developed, originally, in cooperation with the IMS
> Global[5] and are now maintained on [6].
>
>
> "maintained on [6]" does give me pause.  [6] does not state a maintenance
> policy and refers to an issue tracker that uses the pronoun "I" in many
> places, including its Resolved Issues section, and was last modified on 5
> January 2018.  The parent page (WebSchemas) is explicitly disclaimed as
> "left primarily for historical record".  Is this in fact the authoritative
> place for maintaining the current accessibility vocabulary?
>
> The reason of this somewhat weird setting in [2] is to avoid normatively
> referring to schema.org <http://schema.org>.
>
>
> If the WG believes such a normative reference is what the Web needs, it
> should not shy away from stating that.
>
>   Actually, the accessibility spec has an earlier version published at the
> ISO, and in ISO land it was a clear no-no to do so. However, W3C is meant
> to be more flexible and therefore the question does arise. However, our
> document on normative references[7] is not 100% clear cut for me.
> Hence this mail: does W3C has an official position as for a normative
> reference to schema.org <http://schema.org> terms?
>
>
> In this, as in many things, if the WG is able to obtain a clear and
> authoritative statement on the stability of the parts it wants to
> normatively reference, the organization (or community) who "owns" that
> stability, and the open process by which the referenced material is
> maintained, that is important to the Director's consideration.
>
> Specifically, is it possible to simplify [1] and make a clear reference to
> schema.org <http://schema.org> instead of the hand-weaving approach we
> have there currently? In case of a positive answer, can we, possibly, add a
> reference to schema.org <http://schema.org> in [7] just as we do with the
> WhatWG?
>
>
> It depends on the answers to the questions above (and maybe other
> questions that could arise) :)
>
> -Ralph
>
> Thanks for your help
> Ivan
> [1] https://www.w3.org/TR/epub-a11y-11/ <
> https://www.w3.org/TR/epub-a11y-11/>
> [2] https://www.w3.org/TR/epub-a11y-11/#sec-disc-package <
> https://www.w3.org/TR/epub-a11y-11/#sec-disc-package>
> [3] https://www.w3.org/TR/epub-a11y-tech-11/#meta-002 <
> https://www.w3.org/TR/epub-a11y-11/#sec-disc-package>
> [4] https://schema.org/accessMode <https://schema.org/accessMode>
> [5] http://www.imsglobal.org/activity/accessibility <
> http://www.imsglobal.org/activity/accessibility>
> [6] https://www.w3.org/wiki/WebSchemas/Accessibility <
> https://www.w3.org/wiki/WebSchemas/Accessibility>
> [7] https://www.w3.org/2013/09/normative-references <
> https://www.w3.org/2013/09/normative-references>
> ----
> Ivan Herman, W3C
> Home: http://www.w3.org/People/Ivan/ <http://www.w3.org/People/Ivan/>
> mobile: +33 6 52 46 00 43
> ORCID ID: https://orcid.org/0000-0003-0782-2704 <
> https://orcid.org/0000-0003-0782-2704>
>
>
>
> ----
> Ivan Herman, W3C
> Home: http://www.w3.org/People/Ivan/
> mobile: +33 6 52 46 00 43
> ORCID ID: https://orcid.org/0000-0003-0782-2704
>
>