Re: Readium as reference implementation in W3C

Folks,

While this thread is discussing the differences between user-agents and
browsers, I thought it would be of interest to share a little experiment I
did this morning. As I mentioned in NYC (and have done so before) the Test
Suite actually tests the browser in almost half the tests.  As a reading
system (user agent) we (Readium) just delegate content to the browser
engine.  I would argue that in those cases that those obviously are items
not in the DPUB WG purview and can be skipped.

So I went through a spreadsheet we use for recording the results of testing
the Readium Chrome app and the CloudReader.  For each feature which, at
least for Readium, really just tests the browser, I entered a 0 (I.e. fail)
for each test that actually invokes, at least partially, the reading system
I gave a 1 (I.e. pass).  The spreadsheet has a summary at the top.  As you
can see , only 55% of the tests actually test the reading system (less than
half if you exclude accessibility).

Of course, some of this is a judgement call since a RS could presumably
interfere or block the browser from executing in some cases, but, at least
for Readium, this spreadsheet is fairly accurate.

To respond to part of thread below, bear in mind that the Readium
CloudReader (the browser-based version of ReadiumJS) runs in virtually every
browser while there is no requirement that a version of the SDK need use a
specific browser engine, but that is usually dictated by the OS, especially
on mobile.  With Readium-2 the situation is even more fluid.  Point is, we
are probably going to have to carefully weigh these issues on a case by case
basis.

Finally, I could be wrong, but I believe that RMSDK was/is the last known
reading system that DOESN¹T use a browser engine.  If someone knows of one
that isn¹t I would be curious to hear about it.

Best,
Ric

From:  Tzviya Siegman <tsiegman@wiley.com>
Date:  Monday, July 10, 2017 at 9:24 AM
To:  Avneesh Singh <avneesh.sg@gmail.com>, Laurent Le Meur
<laurent.lemeur@edrlab.org>
Cc:  Garth Conboy <garth@google.com>, Ivan Herman <ivan@w3.org>, Daniel Weck
<daniel.weck@gmail.com>, George Kerscher <kerscher@montana.com>, Ric Wright
<rkwright@geofx.com>
Subject:  RE: Readium as reference implementation in W3C

Hi all,
 
A discussion like this would be great to have on the larger
public-publ-wg@w3.org email list.
 
For the purposes of WP/PWP/EPUB 4, user agents do not need to be browsers.
What we need to assess is whether the engines are distinct. The discussion
we had at our F2F was about whether two browsers using Blink, for example,
count as two implementations or one. The same question will be asked of
Readium. 
 
Tzviya
 

Tzviya Siegman
Information Standards Lead
Wiley
201-748-6884
tsiegman@wiley.com <mailto:tsiegman@wiley.com>
 

From: Avneesh Singh [mailto:avneesh.sg@gmail.com]
Sent: Friday, July 07, 2017 2:03 PM
To: Laurent Le Meur <laurent.lemeur@edrlab.org>
Cc: Siegman, Tzviya - Hoboken <tsiegman@wiley.com>; Garth Conboy
<garth@google.com>; Ivan Herman <ivan@w3.org>; Daniel Weck
<daniel.weck@gmail.com>; George Kerscher <kerscher@montana.com>; Ric Wright
<rkwright@geofx.com>
Subject: Re: Readium as reference implementation in W3C
 

³On Readium JS  (Readium Cloud reader and Readium Chrome app),  if the
webview in use supports aria-details, we could say that Readium
automatically gains supports for it.²

 

Hi Laurent, 

Indeed this is why I aslked Joanie if Readium will be considered as separate
reference implementation from Google Chrome which implements aria-details.

 

Actually my email was beyond accessibility. It is to get clarity, if Readium
is considered as reference implementation from PWG work. It may make things
somewhat easy for PWP/EPUB 4.

 

With regards

Avneesh

From: Laurent Le Meur

Sent: Friday, July 7, 2017 19:02

To: Avneesh Singh 

Cc: Tzviya - Hoboken Siegman ; Garth Conboy ; Ivan Herman ; Daniel Weck ;
George Kerscher ; Ric Wright

Subject: Re: Readium as reference implementation in W3C

 

Hi Avneesh,  

 

As an aside, Readium, as a set of open-source community-based projects, is
mostly going where its developers move it.

EDRLab is willing to help on the accessibility side, this is in our charter.
About the aria-details attribute, I don't know what we could do, precisely.

 

Readium SDK is a reading engine, therefore the html page is fetched but not
rendered. It's up to the desktop or mobile app to render aria-details
correclty.

 

On Readium JS  (Readium Cloud reader and Readium Chrome app),  if the
webview in use supports aria-details, we could say that Readium
automatically gains supports for it. But maybe this is not the UX visually
impaired people would like. Most are using a screen-reader, and in this case
it's up to the screen-reader to render the details. Other people may want a
visual interaction, adapted to their vision; but is such a case, overloading
the webview UX would be a pain. And to do what?

 

Note that we will make our best to create an accessible Readium Desktop app
from this end of this year (the first version, for November, will have no
specific a11y features). If we can have a clear set of features to implement
(easily) by then, this will guide our work.

 

Laurent

 
> 
> Le 7 juil. 2017 à 08:08, Avneesh Singh <avneesh.sg@gmail.com> a écrit :
> 
>  
> 
> Interesting discussion in one more thread.
> In Lisbon TPAC, Ivan mentioned that Readium can be a valid reference
> implementation for Rec deliverable's.
> 
> So, did we formalize this? or is this something that we should take up in next
> calls of PWG.
> 
> 
> With regards
> Avneesh
> -----Original Message----- From: Joanmarie Diggs
> Sent: Friday, July 7, 2017 11:29
> To: George Kerscher ; 'Avneesh Singh'
> Cc: 'Tzviya - Hoboken Siegman' ; 'Matt-Garrish' ; 'Daniel Weck' ; 'Charles
> LaPierre'
> Subject: Re: aria-details at risk
> 
> Hi George.
> 
> I can look into how non-browser implementations can be added to the test
> results. In the meantime, as I indicated in my response to Avneesh, we
> are not marking aria-details at risk.
> 
> --joanie
> 
> On 07/07/2017 01:50 AM, George Kerscher wrote:
> 
>> 
>> Hi Joanie,
>> 
>> 
>> 
>> Now that there is the  Publishing at the W3C initiative,
>> 
>> https://www.w3.org/publishing/
>> 
>> an implementation in a ³reading system,² such as Readium,  I suggest
>> that this should be considered equal to a browser implementation. As
>> Avneesh indicates, we could ask/direct Readium to implement support
>> ASAP. We can also get the word out that other reading systems implement
>> support as well.
>> 
>> 
>> 
>> Best
>> 
>> George
>> 
>> 
>> 
>> *From:*Avneesh Singh [mailto:avneesh.sg@gmail.com]
>> *Sent:* Thursday, July 6, 2017 9:55 PM
>> *To:* jdiggs@igalia.com
>> *Cc:* Tzviya - Hoboken Siegman <tsiegman@wiley.com>; George Kerscher
>> <kerscher@montana.com>; Matt-Garrish <matt.garrish@gmail.com>; Daniel
>> Weck <daniel.weck@gmail.com>; 'Charles LaPierre' <charlesl@benetech.org>
>> *Subject:* aria-details at risk
>> 
>> 
>> 
>> Hi Joanie,
>> 
>> 
>> 
>> Tzviya alerted me that aria-details is at risk for ARIA 1.1.
>> 
>> We would like to know how we can help in getting it through CR. It looks
>> that you are tracking browser implementations, and it is supported only
>> in Google Chrome at this time.
>> 
>> We have done the following work for aria-details.
>> 
>> 1. It is implemented in one of DAISY Consortium production tools
>> 
>> http://www.daisy.org/tobi/change-log
>> 
>> 2. We are working on sample content, it will be implemented in DIAGRAM
>> image description sample book finally.
>> 
>> The work in progress is at:
>> https://rawgit.com/daisy/aria-details/master/index.html
>> 
>> 3. We can request Readium project to implement it, in case it is
>> considered as a separate reference implementation.
>> 
>> http://readium.org/
>> 
>> 
>> 
>> With regards
>> 
>> Avneesh Singh
>> 
>> COO, DAISY Consortium