Re: Best citation format for accessibility

IMO there is no real universality about the mapping of DocBook to HTML5
that has been accomplished, for EPUB's use of HTML5, in EDUPUB. It is just
a particular microformat, just as DocBook itself is just a particular XML
schema. The EDUPUB work was based on submissions of prior work on
DocBook-to-HTML mapping by Pearson (PXE) and O'Reily (HTMLBook for Atlas),
so it has some demonstrated real-world utility, but the scientific
community clearly has additional requirements that were not central to the
education and technical publishing fields that Pearson and O'Reilly were
primarily concerned with, and wherein DocBook was already being somewhat
widely used.

So I think it'd be just fine for folks to work on e.g. a JATS flavor
microformat of HTML, and in some sense it is inevitable. But at the same
time I would argue that this group shouldn't get hung up on trying to
define "one microformat to rule them all", precisely because the needs of
different segments of publishing/communication re: semantics are different
enough to preclude full unification. This EDUPUB semantic profile of EPUB 3
was developed specifically for EDU segment, and while we do see broader
uses cases for it, it's not something we are likely to consider suitable
for e.g. EPUB 3.1.

This further implies that for both machine analysis and a11y use cases,
"polyglot" processing will likely be required. There won't be just one
semantically-enhanced flavor of HTML out in the world, but a number of
them. I think therefore that the best that this group can do is to develop
the broader building blocks and best practices, and if we do have some
"good enough" semantic enhancements that everyone adopts, that's upside
gravy not something we have to depend on happening for sure.

--Bill


On Tue, Sep 22, 2015 at 9:43 AM, Bill Kasdorf <bkasdorf@apexcovantage.com>
wrote:

> Yup. My initial comments were more about the machine consumption. My later
> comments were in response to Tzviya's comment about users needing to
> actually read and interpret the citation. Both are useful.
>
> And yes, for machine consumption, capturing all the information that is
> already in JATS is what I was suggesting.
>
> As for a link to the JATS data, that's an interesting thought. One recent
> addition to JATS is a NISO standard called ALI, Access and License
> Indicators, which was created for hybid journals that have both open access
> and subscription access. It consists of two elements, one of which provides
> a stable URI to where the license terms are available. (The other simply
> says whether the article is free to read or not.) Your suggestion is sort
> of a variant on that. (Please don't shoot me for bringing up the
> open-or-not issue.)
>
> But frankly, if the purpose is just getting to the cited resource, then
> rather than a link to the JATS info, it seems more straightforward to just
> implement the link to the cited resource that is usually already in that
> citation (having been put there based on the JATS markup in the citation
> and in the metadata of the cited article).
>
> --Bill
>
> -----Original Message-----
> From: Leonard Rosenthol [mailto:lrosenth@adobe.com]
> Sent: Tuesday, September 22, 2015 12:31 PM
> To: Bill Kasdorf; Siegman, Tzviya - Hoboken; Robin Berjon;
> public-digipub-ig@w3.org
> Subject: Re: Best citation format for accessibility
>
> Bill - I agree with you here concerning a PERSON who is consuming this
> content.
>
> What is missing is providing the necessary information for MACHINE
> consumption (and understanding) of the content.  It would seem to me
> (without having any understanding of this area) that I would want to be
> able to encode all of the JATS information directly in the format (HTML).
> Certainly, that would seem like the best approach.  Alternatively, and I
> don’t know if this makes sense (though it sounds like it does), I’d like a
> link/URI to that JATS data.  Sort of a “aria-describedat” if you will.
>
> Leonard
>
>
>
>
>
> On 9/22/15, 12:23 PM, "Bill Kasdorf" <bkasdorf@apexcovantage.com> wrote:
>
> >Exactly, I meant to mention that. Ideally, the citation in the text is
> linked to the full citation at the end of the article, all within the
> document.
> >
> >I think the point is that in the scholarly literature (and here I am only
> talking about that), the act of _reading_ a citation is for two main
> purposes:
> >
> >--The main purpose is for other authors to see if they have been cited by
> this author. ;-)
> >
> >--The next purpose is to see if the cited resource is something they need
> to get, or get to.
> >
> >They don't actually need the detailed granular semantics of the citation
> to accomplish the latter. They see the link (the DOI-based link in the full
> citation) and click on it and get to where they're going.
> >
> >But your point is well taken. For AT, a user wants to be able to do those
> things, and it ain't easy. I still think there's potential in the fact that
> the citation does in fact have mostly sufficient markup. For a given
> citation, they will mainly want to know "who are the authors?", "what's the
> title of the article?", "where was it published?", "when was it
> published?", and then "where do I click to get to it?" (That's my personal
> opinion; others may have other opinions.) Since the semantics that
> designate those things are usually (almost always) in the JATS markup of
> the citation, I think there's potential to look at that and say "how do we
> take that and make sure it works for AT?" This might involve a bit of
> refinement to the proposed ARIA structural semantics for citations. But
> maybe what's there is sufficient, and it's a matter of making sure that the
> places where these citations live online--in hosting services like
> HighWire, Atypon, and PT, and in the sites of big publishers like Wiley,
> Springer, and Elsevier, at the present time; again, I realize that in the
> future we want more than that) are wired to connect the dots from the JATS
> markup they get to what is needed to make the citation accessible.
> >
> >I'm not saying I have a solution yet, I'm just suggesting what might be a
> productive path forward, and might address this accessibility need in the
> near term for 90+% of the scholarly citations that are out there.
> >
> >--Bill K
> >
> >-----Original Message-----
> >From: Siegman, Tzviya - Hoboken [mailto:tsiegman@wiley.com]
> >Sent: Tuesday, September 22, 2015 12:06 PM
> >To: Bill Kasdorf; Robin Berjon; public-digipub-ig@w3.org
> >Subject: RE: Best citation format for accessibility
> >
> >The reference to the citation and the citation are rather linked.
> >
> >As you noted, Bill, JATS is robust, but how does that get to a user?
> >
> >As a sited (cited?) reader, I find citations to be mind-blowingly awful
> to navigate. Citations are often lists within footnotes. Imagine the number
> of AT hops required!
> >
> >Tzviya Siegman
> >Digital Book Standards & Capabilities Lead Wiley
> >201-748-6884
> >tsiegman@wiley.com
> >
> >
> >-----Original Message-----
> >From: Bill Kasdorf [mailto:bkasdorf@apexcovantage.com]
> >Sent: Tuesday, September 22, 2015 12:00 PM
> >To: Siegman, Tzviya - Hoboken; Robin Berjon; public-digipub-ig@w3.org
> >Subject: RE: Best citation format for accessibility
> >
> >Just pointing out that Tzviya is talking (at least partly) about
> something different than I was talking about.
> >
> >She is talking, in the context of Name/Date etc. ("Melville, 1851"),
> about how the citation is expressed in text, typically referencing a full
> citation provided in a bibliography or reference list (typically in a
> section at the end of the paper, in backmatter). She's exactly right, as
> usual. There are lots of different ways of doing that in-text citation, and
> many of those don't give you much to work with. But they aren't typically
> designed to take you to the cited thing, they are designed to take you to
> the citation of the cited thing.
> >
> >I was talking about that full citation in the bibliography or reference
> list. Those are typically very richly tagged (at least _adequately_ tagged)
> to enable CrossRef resolution, which I think would provide useful semantics
> for accessibility. They work extremely well for citations of journal
> articles; less well but better all the time for citations to books,
> chapters, conference proceedings, standards, websites, etc.
> >
> >--Bill Kasdorf
> >
> >-----Original Message-----
> >From: Siegman, Tzviya - Hoboken [mailto:tsiegman@wiley.com]
> >Sent: Tuesday, September 22, 2015 11:33 AM
> >To: Robin Berjon; public-digipub-ig@w3.org
> >Subject: RE: Best citation format for accessibility
> >
> >Hi Robin,
> >
> >Great to hear from you.
> >
> >There is a great lack of consensus on best practices for citation in
> general. Harvard vs AMA vs Vancouver, and that does not include the variety
> of HTML citations. They all include essentially the same information in a
> different sequence.
> >
> >I turn the question around to you. What is missing? Should citations be
> chunked elements that a user can tab through? If AT can pick up on existing
> ontologies we can do this now using resources like BIBO [1],  CITO [2], and
> others. (This might not be easy, but it's better than <span
> class="surname">)
> >
> >One pain point I see is the Name Date method of citation, which refers
> the reader to the citation by use of the authors surname and year of
> publication. For example, a reference to Moby Dick would be (Melville,
> 1851). Multiple references to the same work would use the same reference.
> When digital, these references are usually links.  I think this method of
> linking violates WCAG unless one is really careful.
> >
> >Should we loop in WAI?
> >
> >[1] http://bibliontology.com/
> >[2] http://www.essepuntato.it/lode/http://purl.org/spar/cito
> >
> >Tzviya Siegman
> >Digital Book Standards & Capabilities Lead Wiley
> >201-748-6884
> >tsiegman@wiley.com
> >
> >
> >-----Original Message-----
> >From: Robin Berjon [mailto:robin@berjon.com]
> >Sent: Tuesday, September 22, 2015 10:41 AM
> >To: public-digipub-ig@w3.org
> >Subject: Best citation format for accessibility
> >
> >Hi,
> >
> >citations in scholarly publishing have a long history of at-time
> acrimonious disagreement over the exact format one should set them in.
> >There can be long arguments about the how and why of some specific
> detail, but these are all about visual presentation. I have yet to hear
> someone discuss the best format to use for the *content*, when in digital
> form, such that it is most accessible.
> >
> >By applying some technology, we can reformat a citation for visual
> rendering. We can even make citation formatting follow readers'
> >preferences rather than publishers'. But when doing so the HTML-level
> encoding of the citations should be optimised for semantic, non-visual
> access.
> >
> >So my question is: has anyone given thought to what the best order of
> content and best markup practices would be for optimally accessible
> citations?
> >
> >Thanks!
> >
> >--
> >Robin Berjon - http://berjon.com/ - @robinberjon
> >
>



-- 

Bill McCoy
Executive Director
International Digital Publishing Forum (IDPF)
email: bmccoy@idpf.org
mobile: +1 206 353 0233