RE: Running headers/footers / new version of DPUB-ARIA?

Hi,
 
There is a Word to EPUB proof of concept. Here is what it does with page numbers:
 
Page numbers in EPUB are requested again and again, so we include these in our WordToEPUB proof of concept. Since they are not included in the XML (docx) file, they are generated in a running instance of Word, moving through to the start of each page programmatically (using Word interop), and adding the proper markup. The document may have adjusted page numbers (perhaps they start at page 25), and so this is accounted for.
 
A page list is generated in thenavdoc, the source of page numbers is noted in the EPUB metadata, and the presence of page numbers noted in the accessibility metadata.
 
The technique works well in a wide range of reading systems on different platforms.
 
Hope this helps.

Best
George
 
 
From: Brady Duga <duga@google.com> 
Sent: Thursday, August 22, 2019 9:33 AM
To: Aaron Leventhal <aleventhal@google.com>
Cc: George Kerscher <kerscher@montana.com>; Avneesh Singh <avneesh.sg@gmail.com>; White, Jason J <jjwhite@ets.org>; Leonard Rosenthol <lrosenth@adobe.com>; W3C Publishing Working Group <public-publ-wg@w3.org>; Joanmarie Diggs <jdiggs@igalia.com>
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
Docs does in fact have native epub export (File -> Download -> EPUB Publication), largely driven by Garth Conboy (also a co-chair of this group). However, there is currently no facility in epub to do dynamic page numbering, though of course implementations may generate them for their UI. In the past, dynamic page numbering that changes with reflow has been experimented with, but the community has largely moved to fixed page numbers. This allows a user to understand their location after eg text zoom (page 250 is always 50% of a 500 page book). The other common use is to share location information succinctly and precisely, eg "class, turn to page 250". It is unclear to me how important having a generated page number is. There is also no particularly good way to do running page-header/foot in epub, though a hacky facility to do it did exist at one point. Currently, running heads are dropped in Docs to epub export. Future work in this area will likely be done in collaboration with CSS, assuming we actually decide to do it. 
 
On Thu, Aug 22, 2019 at 8:01 AM Aaron Leventhal <aleventhal@google.com <mailto:aleventhal@google.com> > wrote:
George, sounds like you are saying that if Docs is to export to EPUB (I don't think it currently does, but maybe there's an extension?) -- then there would be a set of requirements to ensure good quality navigation.
 
Aaron
 
On Thu, Aug 22, 2019 at 10:48 AM George Kerscher <kerscher@montana.com <mailto:kerscher@montana.com> > wrote:
Hello,
OK, when using  Google Docs and using the publish this document, the spot of the page break at the top must contain the  doc-page-break attribute. Also, when generating the EPUB version, this is the spot 
That must be linked Along with the number that must be in the page list in the navdoc.
In this way, people using the print version and the EPUB version can be on the same page, grin.
 
Best
George
 
From: Aaron Leventhal <aleventhal@google.com <mailto:aleventhal@google.com> > 
Sent: Thursday, August 22, 2019 7:08 AM
To: Avneesh Singh <avneesh.sg@gmail.com <mailto:avneesh.sg@gmail.com> >
Cc: White, Jason J <jjwhite@ets.org <mailto:jjwhite@ets.org> >; George Kerscher <kerscher@montana.com <mailto:kerscher@montana.com> >; Leonard Rosenthol <lrosenth@adobe.com <mailto:lrosenth@adobe.com> >; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> ; Joanmarie Diggs <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
Avneesh, Google Docs has a "publish this document" feature, as well as a mode where users can read, but not edit a document. In the published case, there is no caret navigation built in. Being able to skip over running page content should work in any of these modes.
 
When zooming in the browser, I would not currently expect pagination to change at all. However, if such a zoom and reflow feature did exist, and repagination did occur, I imagine that the user would want to be able to refer to original page numbers, similarly to Braille print page indicators. I'm not sure how this would be handled, but I suppose it's worth discussing how the role doc-pagenum would be used. Let's think it out. Personally I don't think it should be required for the doc-pagenum role to be inside of a doc-pageheader or doc-pagefooter, but that it could optionally be.  I don't see it as an issue to have the browser/AT combo work with either scenario.
 
Aaron
 
On Thu, Aug 22, 2019 at 7:53 AM Avneesh Singh <avneesh.sg@gmail.com <mailto:avneesh.sg@gmail.com> > wrote:
Clarifying it further. My example is from publishing perspective i.e. mostly read only documents that are published.
If we are looking at this from purely word processor perspective, then I wonder if this is a publishing related issue?
 
With regards
Avneesh
Sent: Thursday, August 22, 2019 17:19
To: White, Jason J ; George Kerscher ; 'Aaron Leventhal' 
Cc: 'Leonard Rosenthol' ; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org>  ; 'Joanmarie Diggs' 
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
 
I think that I am able to get the expected functionality, at the same time I am still confused with placing page numbers in footers or headers. 
Explaining it in step wise manner
- The text will reflow but the headers and footers are expected to remain constant.
- For example, if the document has footers after approximately 20 lines and the user zooms to double size, then there will be one footer after each 7 or 8 lines.
- This means that page breaks will also flow down.
 
Now, how can one mark persistent page number inside header or footer. In such a case the page number in footer will become reading system responsibility which will change dynamically instead of being responsibility of content.
 
If we want to have static page numbers in header or footers then it should be a fixed layout,
or the header / footer should also reflow when user zooms the document.
 
With regards
Avneesh
From: White, Jason J 
Sent: Thursday, August 22, 2019 1:35
To: George Kerscher ; 'Aaron Leventhal' ; 'Avneesh Singh' 
Cc: 'Leonard Rosenthol' ; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org>  ; 'Joanmarie Diggs' 
Subject: RE: Running headers/footers / new version of DPUB-ARIA?
 
Perhaps the screen readers should offer a means of querying the header/footer information. However, that’s outside the scope of ARIA and a matter for each implementation. This would be especially useful in a word processor, as the user needs to be able to read and edit it. Another option would be a simple “go to header/footer” navigation command in the application itself.
 
I’m not sure that I fully understand George’s question about export to EPUB. When the text is reflowed, the pagination may change, but the header/footer text often wouldn’t. A good example would be printing to different page sizes or font sizes. If the text still fits, then presumably it should simply be printed wherever the new page breaks lie after styles are applied to the document. It’s very much the same as what happens if you make changes to a document in a word processor. If there are no explicit page breaks in the markup, then the page header/footer should be able to be applied to an entire div/section of the document:
 
<section role=”doc-chapter”>
<!—Applies to all pages in this section -->
<div role=”doc-pageheader”>[…]</div>
[…]
</section>
 
I’m not in a position to answer Aaron’s earlier question about the schedule for specifying new ARIA features for these use cases. I think that’s a question for the Chairs, or for a group meeting.
 
From: George Kerscher <kerscher@montana.com <mailto:kerscher@montana.com> > 
Sent: Wednesday, August 21, 2019 12:50 PM
To: 'Aaron Leventhal' <aleventhal@google.com <mailto:aleventhal@google.com> >; 'Avneesh Singh' <avneesh.sg@gmail.com <mailto:avneesh.sg@gmail.com> >
Cc: 'Leonard Rosenthol' <lrosenth@adobe.com <mailto:lrosenth@adobe.com> >; White, Jason J <jjwhite@ets.org <mailto:jjwhite@ets.org> >; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> ; 'Joanmarie Diggs' <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: RE: Running headers/footers / new version of DPUB-ARIA?
 
Hi,
 
A few items here:
 
I totally agree that while creating content as in GoogleDocs, it is perfectly fine to have running footers and Headers. Even better if I, using my screen reader, do not need to encounter them.
 
The export to EPUB would not put this in the reflowed content, right?
 
Page numbers are one of those things we have called “skippable” structures. In an audio book, you may want to go to page 34 and hear that this is page 34, but in continuous reading, you want to skip (filter is your term) over it.  However,  It is always good to stop and find out where you are.
 
Seems that this agreement.
 
Best
George
 
From: Aaron Leventhal <aleventhal@google.com <mailto:aleventhal@google.com> > 
Sent: Wednesday, August 21, 2019 8:46 AM
To: Avneesh Singh <avneesh.sg@gmail.com <mailto:avneesh.sg@gmail.com> >
Cc: Leonard Rosenthol <lrosenth@adobe.com <mailto:lrosenth@adobe.com> >; White, Jason J <jjwhite@ets.org <mailto:jjwhite@ets.org> >; George Kerscher <kerscher@montana.com <mailto:kerscher@montana.com> >; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> ; Joanmarie Diggs <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
George -- we're in agreement. This will help paginated content just seem like one continuous document when you're reading. Basically, we're trying to make it so that you don't see this stuff when you're word processing unless you want to. By having the word processor (like Google Docs) mark these up, the AT can decide to filter those things out, which they can't do right now. See Glen Gordon's comments in the bug. This will allow them to have similar functionality for this in Google Docs as they do on MS word. Where are we not aligned here?
 
Avneesh -- DPUB is very close to being able to help us have an awesome experience in Google Docs. Makes sense as a use case, right?
 
Jason -- can you give an example? I think we want 2 things from the page number -- to be able to report it when the user wants it, and to be able to filter it out when the user just wants to read continuously (similar to filtering out the page header).
 
On Wed, Aug 21, 2019 at 10:38 AM Avneesh Singh <avneesh.sg@gmail.com <mailto:avneesh.sg@gmail.com> > wrote:
I think that the confusion comes from the design. Most of the DPUB aria roles are from the perspective of a refloable document like HTML and EPUB 3. On the other hand the discussion in this thread looks more oriented towards somewhat like a fixed layout document. 
 
With regards
Avneesh
From: Aaron Leventhal 
Sent: Wednesday, August 21, 2019 19:56
To: Avneesh Singh 
Cc: Leonard Rosenthol ; White, Jason J ; George Kerscher ; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org>  ; Joanmarie Diggs 
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
Thanks, I didn't realize that's where the page number would currently go. I'm not sure it's ideal. 
<span id="pg04" role="doc-pagebreak" title="4"/>
A page break line would likely show up as a horizontal line. Under the horizontal line is a header, then adjacent to that comes the page number. How would that be done in markup?
We don't want 2 page breaks, and the page break and page number aren't in the same place.
<hr role="doc-pagebreak">
<div role="doc-pageheader">
  My header <span role="doc-pagenumber">4</span>
</div>
 
Additionally, in the page break example, the page number usage is in the title attribute, which would not be visible text. Although it's not explained, t's not clear to me from the example or spec text that browsers/ATs could expect a page number inside a child text node.
 
IMO we should have a separate role for page numbers, or put them inside the header or footer objects, but putting them as part of the page break to me will cause a problem with the above example.
Aaron
 
On Wed, Aug 21, 2019 at 10:17 AM Avneesh Singh <avneesh.sg@gmail.com <mailto:avneesh.sg@gmail.com> > wrote:
doc-pagebreak already exists in DPUB aria roles for locating the page marks. They can have a visible page number as well as invisible page number.
 
 
With regards
Avneesh
From: Aaron Leventhal 
Sent: Wednesday, August 21, 2019 19:12
To: Leonard Rosenthol 
Cc: White, Jason J ; George Kerscher ; public-publ-wg@w3.org <mailto:public-publ-wg@w3.org>  ; Joanmarie Diggs 
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
Thanks Leonard, don't we want to be able to differentiate the headers from the running headers? It's the repeated running headers that users want to skip over.  
The headers that are part of the content seem different, like a <header> in HTML.
 
On Wed, Aug 21, 2019 at 9:35 AM Leonard Rosenthol <lrosenth@adobe.com <mailto:lrosenth@adobe.com> > wrote:
In PDF, we treat page numbers as a separate thing from the header/trailer (and yes, you can have the former inside the latter).  We also have a “class” for bates numbers and a few other vertical-specific elements.   We have found that it is important to be able to unique identify them.
 
Headers and Footers aren’t necessary tied to pages – you could have a section with a header or footer as well, so perhaps just `doc-header` and `doc-footer` (and `doc-pagenum`).
 
Leonard
 
From: Aaron Leventhal <aleventhal@google.com <mailto:aleventhal@google.com> >
Date: Wednesday, August 21, 2019 at 8:55 AM
To: "White, Jason J" <jjwhite@ets.org <mailto:jjwhite@ets.org> >
Cc: Leonard Rosenthol <lrosenth@adobe.com <mailto:lrosenth@adobe.com> >, "kerscher@montana.com <mailto:kerscher@montana.com> " <kerscher@montana.com <mailto:kerscher@montana.com> >, "public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> " <public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> >, Joanmarie Diggs <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
 
Sounds good.  
- How about doc-pageheader and doc-pagefooter?
- Possible definition: repeated section of text that appears at the top/bottom of pages in a document, potentially containing a title, page number or other information
- Should the page number get its own role, e.g. doc-pagenum? Should it exist within the doc-pageheader/doc-page-footer?
- Should the new roles inherit from contentinfo?
 
On Wed, Aug 21, 2019 at 8:05 AM White, Jason J <jjwhite@ets.org <mailto:jjwhite@ets.org> > wrote:
If there’s already a mapping to accessibility APIs via PDF readers, then we could use it here too.
 
Screen readers could offer a command to query this information, if desired, just as there is for window titles.
 
From: Leonard Rosenthol <lrosenth@adobe.com <mailto:lrosenth@adobe.com> >
Date: Tuesday, August 20, 2019 at 21:45
To: George Kerscher <kerscher@montana.com <mailto:kerscher@montana.com> >, 'Aaron Leventhal' <aleventhal@google.com <mailto:aleventhal@google.com> >, "public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> " <public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> >
Cc: 'Joanmarie Diggs' <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: Re: Running headers/footers / new version of DPUB-ARIA?
Resent-From: <public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> >
Resent-Date: Tuesday, August 20, 2019 at 21:45
 
FWIW: Headers & Footers can be properly identified in PDFs and exposed accordingly to AT (so that they aren’t read multiple times).  In fact, PDF/UA (the PDF standard for Universal Accessibility) calls out their usage.
 
Leonard
 
From: "kerscher@montana.com <mailto:kerscher@montana.com> " <kerscher@montana.com <mailto:kerscher@montana.com> >
Date: Tuesday, August 20, 2019 at 4:35 PM
To: 'Aaron Leventhal' <aleventhal@google.com <mailto:aleventhal@google.com> >, "public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> " <public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> >
Cc: 'Joanmarie Diggs' <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: RE: Running headers/footers / new version of DPUB-ARIA?
Resent-From: <public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> >
Resent-Date: Tuesday, August 20, 2019 at 4:35 PM
 
Hello,
 
This is just my opinion. Running headers and footers are very intrusive while reading. I encounter these many times in PDF documents. The same information is constantly repeated. Fortunately, I do not encounter this while reading EPUB. I can understand that it might be useful for the Reading System to provide “where am I?” information as one is reading, but including this over-and-over again in the content would be horrible.
I am reading using Assistive Technology (AT) and the TTS reads to me. I imagine this would also be painful if a person was using the read aloud function in Reading Systems.
 
I could see the same problem in an audio book. Nobody would want to hear those running heads and footers in the audio book.
 
My $.02
 
Best
George
 
Best
George
 
 
From: Aaron Leventhal <aleventhal@google.com <mailto:aleventhal@google.com> > 
Sent: Tuesday, August 20, 2019 11:37 AM
To: public-publ-wg@w3.org <mailto:public-publ-wg@w3.org> 
Cc: Joanmarie Diggs <jdiggs@igalia.com <mailto:jdiggs@igalia.com> >
Subject: Running headers/footers / new version of DPUB-ARIA?
 
Hi all,
 
I'm wondering about the plans for doing DPUB-ARIA 2.0 as described under the deliverables:
https://www.w3.org/2017/04/publ-wg-charter/#deliverables <https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.w3.org%2F2017%2F04%2Fpubl-wg-charter%2F%23deliverables&data=02%7C01%7Cjjwhite%40ets.org%7Cba59d4f6e6f044acfd9908d726579a04%7C0ba6e9b760b34fae92f37e6ddd9e9b65%7C0%7C0%7C637020030007316600&sdata=PTB8VcQ%2FW8W5fG01aBOpUEjbgqKo1vPm%2FXbLtjQo3LM%3D&reserved=0> 
 
Thanks for what DPUB-ARIA provides so far. Google Docs is going to be using it  to expose semantics for online word processing. We're collaborating with AT vendors and other developers of online word processors as well.
 
One gap is running headers and footers: https://github.com/w3c/dpub-aria/issues/10 <https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fw3c%2Fdpub-aria%2Fissues%2F10&data=02%7C01%7Cjjwhite%40ets.org%7Cba59d4f6e6f044acfd9908d726579a04%7C0ba6e9b760b34fae92f37e6ddd9e9b65%7C0%7C0%7C637020030007316600&sdata=e3ypq%2B6xcsXXGoZesjzAX7dFC%2BlLIPu74KxiX0YHuVE%3D&reserved=0>  -- hence the reason for my email and status check :)
 
Thanks for any info,
Aaron
 
 

  _____  

This e-mail and any files transmitted with it may contain privileged or confidential information. It is solely for use by the individual for whom it is intended, even if addressed incorrectly. If you received this e-mail in error, please notify the sender; do not disclose, copy, distribute, or take any action in reliance on the contents of this information; and delete it from your system. Any other use of this e-mail is prohibited.
 
Thank you for your compliance.

  _____  

 

  _____  

This e-mail and any files transmitted with it may contain privileged or confidential information. It is solely for use by the individual for whom it is intended, even if addressed incorrectly. If you received this e-mail in error, please notify the sender; do not disclose, copy, distribute, or take any action in reliance on the contents of this information; and delete it from your system. Any other use of this e-mail is prohibited.
 
Thank you for your compliance.

  _____