Re: Extending DPub with new roles floating boxes and references?

Silvio,

The figure element in HTML will map to the new figure role. I am not sure
why the floating box cannot be represented as a figure role. Also, within
the figure you can have tables, images, etc.

<div role="figure" aria-describedby="foo">
   <div role="table">
    ...
   </div>
</div?
<p id="foo>Figure 1.0 shows the median wage for the population in Sappor
from the ages of 18 to 60</p>

The fact that the box is float has never been a use case that anyone has
provided to us. What is the value of saying it is floating?  You will need
to get at it with a keyboard. Did you want a way to say it is a modal
dialog type of figure?

Rich


Rich Schwerdtfeger



From: Silvio Peroni <silvio.peroni@unibo.it>
To: Richard Schwerdtfeger/Austin/IBM@IBMUS
Cc: <public-dpub-aria@w3.org>
Date: 10/26/2015 05:39 PM
Subject: Re: Extending DPub with new roles floating boxes and
            references?



Hi Richard,

First of all, thanks for your answer. I’ve quickly read the documentation
on Aria 1.1, and I’ve found a very interesting use of the “math” role, that
is very related with the two possible options that we have in RASH for
describing mathematical formulas (mentioned my previous email), i.e., one
with the use of mathml, and the other with the use of an image. I think
I’ll integrate the approach in the version of RASH.

However, I still believe that the roles “figure” and “table” you suggested
are not enough for addressing the scenarios I was talking about. Such roles
maps quite well (and according to the Aria spec too) with the related
existing HTML tags, as indicated by the “Related Concepts” relation. Now,
it is clear that they may express similar (not identical) meaning, since an
element having role “figure” are not the same element expressed by using
the tag “figure”. However, the definitions are pretty close, and, as far as
I understood, basically allow one to express any kind of floating boxes,
i.e., “A perceivable section of content that typically contains a graphical
document, images, code snippets, or example text”. However, what we want to
do in RASH is to have a mechanism to clearly identify all these possible
options, i.e.:
- sections expressing a floating box that contains an image;
- sections expressing a floating box that contains a math formula;
- sections expressing a floating box that contains a table;
- sections expressing a floating box that contains a code listing.

We believe that this could not be done by simply looking at the content of
such sections, rather it should be clearly defined by appropriate roles
which specialise that sections. Now, we use the HTML element “figure” to
identify these kinds of sections, but what we really would like to do is,
thus, to have a clear mechanism (e.g., by means of roles) to say “look,
this element ‘figure’ actually is a floating box for a table, called table
box” or “a floating box for a pure graphical image, called figure box”,
etc. That particular characterisation seems not possible by means of Aria
and DPub.

I was wondering if that would be a possible extension to consider for
future release of DPub, since those roles seem very specific to the
publishing domain to me.

Thanks again for your help with this.
Have a nice day :-)

S.


      Il giorno 26 ott 2015, alle ore 20:55, Richard Schwerdtfeger <
      schwer@us.ibm.com> ha scritto:



      Silvio,

      Aria 1.1 has a new Figure role and a Table role:
      http://rawgit.com/w3c/aria/master/aria/aria.html#document_structure_roles



      Captions are referenced via aria-describedby and this is how HTML 5
      will map the underlying caption to the figure. You can also use
      aria-labelledby for labels on figures, etc.

      We also have a math role:
      http://rawgit.com/w3c/aria/master/aria/aria.html#math


      We don't have a formula role.

      We are discussing how to deal with specialized links in Sapporo this
      week. I am not sure how soon that will be resolved. We don't have a
      role="ref"

      Rich


      Rich Schwerdtfeger

      <graycol.gif>Silvio Peroni ---10/26/2015 12:18:57 PM---Hi all, I'm
      Silvio Peroni from University of Bologna, and I'm one of the
      contributors of RASH, the R

      From: Silvio Peroni <silvio.peroni@unibo.it>
      To: <public-dpub-aria@w3.org>
      Date: 10/26/2015 12:18 PM
      Subject: Extending DPub with new roles floating boxes and references?





      Hi all,

      I'm Silvio Peroni from University of Bologna, and I'm one of the
      contributors of RASH, the Research Article in Simplified HTML format,
      available on github at http://github.com/essepuntato/rash, that has
      been already used in academic events (e.g., SAVE-SD 2015,
      http://cs.unibo.it/save-sd/2015/index.html).

      The new version of RASH that will be released in the next 7 days,
      together with all its tools (visualisation, conversion, enhancement),
      will introduce some important features. The main one, in my opinion,
      is the fact that it is more HTML5-oriented and all the structural
      roles will be now provided by means of the attribute @role according
      to the value specified in:

      [DPub] Digital Publishing WAI-ARIA Module 1.0 – W3C Editor's Draft.
      https://rawgit.com/w3c/aria/master/aria/dpub.html


      I have few questions for you on this topic.


      # Structural roles for floating boxes

      In RASH there exist four different kinds of floating boxes (specified
      through the element "figure"), i.e.: figure box, table box, formula
      box, and listing box. As you can image, these are used to define
      referenceable boxes like it happens in other document-oriented
      languages such as LaTeX, ODT, etc.

      In the version of RASH that will publish soon, these are specified by
      using the following roles (still by means of the attribute @role on
      "figure" elements): "picture", "table", "formula" and "listing",
      e.g.:

      <figure id="figure_1" role="picture">
      <p>
      <img src="img/rash.png" alt="The RASH logo!"/>
      </p>
      <figcaption>Caption of the figure.</figcaption>
      </figure>

      There is the possibility of adding some particular and official role
      in [DPub] that enables one to handle these specific structures?

      Note that the disambiguation based on the content of "figure" can not
      hold sometimes, if, for instance, I want to express similar boxes
      with different kinds of contents. In RASH, that is the case of
      formulas, that can be expressed by means of either MathML or images,
      i.e.:

      <figure id="formula_1" role="formula">
      <p>
      <img src="img/formula.png" alt="[a + b]^260 + [a + b]_i"/>
      </p>
      </figure>

      and

      <figure id="formula_1" role="formula">
      <p>
      <math xmlns="http://www.w3.org/1998/Math/MathML">
      <mrow>
      <msup>
      <mfenced open="[" close="]">
      <mrow>
      <mi>a</mi>
      <mo>+</mo>
      <mi>b</mi>
      </mrow>
      </mfenced>
      <mn>260</mn>
      </msup>
      <mo>+</mo>
      <msub>
      <mfenced open="{" close="}">
      <mrow>
      <mi>a</mi>
      <mo>+</mo>
      <mi>b</mi>
      </mrow>
      </mfenced>
      <mi>i</mi>
      </msub>
      </mrow>
      </math>
      </p>
      </figure>


      # References to document objects (e.g., sections)

      While in [DPub] there are roles for referring to footnotes
      (doc-noteref) and bibliographic entries (doc-biblioref), I've not
      found any role for pointing to other meaningful document components,
      such as sections, figures, formulas, tables and listings. RASH is
      using the ad-hoc role "ref" specified on the element "a", as shown in
      the following excerpt:

      <p>In <a href="#metadata" role="ref"></a> we explain how to specify
      the metadata of the article as appropriate elements within
      <code>head</code>, while the rest of the document mainly pertains
      elements within the element <code>body</code>.</p>

      ...

      <section id="metadata">
      <h1>Metadata</h1>
      <p>...</p>
      </section>

      What such "a" element with "ref" role creates is a specific
      link/reference to an existing object (a section in this case) that
      will be then rendered as, for instance, "Section 2" in the browser.

      This role seems to behave like the other references included in
      [DPub] (i.e., "doc-biblioref" and "doc-noteref"). What about adding a
      generic "ref" for dealing with references to other kinds of objects?

      --------------

      What do you think about these issues?

      Have a nice day :-)

      S.


      ----------------------------------------------------------------------------

      Silvio Peroni, Ph.D.
      Department of Computer Science and Engineering
      University of Bologna, Bologna (Italy)
      Tel: +39 051 2094871
      E-mail: silvio.peroni@unibo.it
      Web: http://www.essepuntato.it

      Twitter: essepuntato





----------------------------------------------------------------------------

Silvio Peroni, Ph.D.
Department of Computer Science and Engineering
University of Bologna, Bologna (Italy)
Tel: +39 051 2094871
E-mail: silvio.peroni@unibo.it
Web: http://www.essepuntato.it

Twitter: essepuntato