RE: ARIA use in HTML other than for accessibility.

> because I was having a terrible time just getting the materials I needed for school in an accessible manner.

This is an excellent use case for why the time came long ago for extensions.

It is highly unlikely that most authors know that they intend to indicate that each chapter is a landmark and each sidebar is an <aside> (default role ="complimentary"). What about the glossary term that pops up when the user taps on the glossary term but also displays as an aggregated list at the end of the chapter?  Should the ARIA mark up only the widget functionality or should there be something telling the user that this is a glossary term and definition embedded in <dl> and associated with the glossary term in the content? Would it be valuable for there to be a role indicating that a region includes assessments instead of simply providing a region with a name "assessment"? This might offer some uniformity in test-taking for standardized tests. These are just a few of the example that publishers face as we attempt to provide accessible content to our users. 

The DAISY Consortium has been publishing for years using the principles that semantics (or if you prefer, inflection) guide usability and accessibility. They developed a structural semantic vocabulary [3] to standardize the structure of written documents.  I know countless people who have relied heavily on DAISY's Digital Talking Book to learn. DAISY and IDPF joined forces to create EPUB 3 [4] as the accessible publication format, with the full expectation that publishers would include proper ARIA markup. EPUB 3 is also the standard for digital publication that the vast majority of publishers and reading systems in the world use. The EPUB structural semantic vocabulary [5] evolved and continues to evolve. One of the amazing things about this vocabulary is that it benefits everyone. I (a publisher) can create one table of contents with extensive CSS that feeds into every reading system and user agents' automated bookmarking tool to generate a table of contents widget, with the help of this vocabulary. It would benefit an even wider audience if there was a clear path forward to map these terms to the accessibility tree.  This is not representing lazy developers, but the varied audience of those developing with ARIA  and taking advantage of its communication with AT. 

I don't think this will be easy, but I think the benefits greatly outweigh the risks.

[1] https://lists.w3.org/Archives/Public/public-html/2015May/0030.html 
[2] https://lists.w3.org/Archives/Public/public-pfwg/2015May/0031.html 
[3] http://www.daisy.org/z3998/2012/vocab/structure/ 
[4] http://idpf.org/epub/30 
[5] http://www.idpf.org/epub/vocab/structure/ 

Tzviya Siegman
Digital Book Standards & Capabilities Lead
Wiley
201-748-6884
tsiegman@wiley.com 


-----Original Message-----
From: Bryan Garaventa [mailto:bryan.garaventa@ssbbartgroup.com] 
Sent: Monday, May 04, 2015 7:34 PM
To: _mallory
Cc: W3C WAI Protocols & Formats
Subject: RE: ARIA use in HTML other than for accessibility.

> Those whom I trust as a developer to tell me accurate things aren't 
> teaching at university. They're working as engineers right now. Most of them are more familiar with the field because they're creating it themselves, at companies.

I agree that this is the case now, however I take what many companies build apart every day, and I can attest from years of doing this that the vast majority of mainstream developers at companies still have a very limited understanding of what ARIA does and how it specifically interacts with ATs, in many ways causing more problems than it solves when applied without this level of understanding.

The power in ARIA is that it directly interfaces with Assistive Technologies. This is also the primary danger of it, because if developers don't take this into account when using it, they can easily break accessibility instead of enhancing it.

A simple example of this is the use of role=alert when applied to a timer field that counts the number of seconds displayed there. It may seem useful, but since it ties into the alert system on the platform operating system, it causes nothing but this information to be constantly repeated no matter what a screen reader user is doing on the computer, effectively hijacking their system.

> Last time I saw "web development" in higher education, it was still espousing <center> and <font> tags and everything in frames. In 2012.

It's true that many old-school learning materials are often out of date given the pace of development that we face, but that's no excuse for not trying to provide better learning materials for those who are trying to learn while still in school.

When provided as E-learning materials, the same materials are available to all developers whether they are enrolled in school or not, thus maximizing exposure and knowledge transfer. These don't have to be books necessarily, but actual interactive online materials that users can learn from, provided preferably by those here at the W3C to add weight to their validity. This goes beyond the scope of simple tutorials however, which is where the background and platform level aspects of ARIA have to be conveyed at the same time.

This way new and learning developers won't have to continually return to W3Schools for this purpose instead.

> You have gone through this before-- was it something most people in the field learned by themselves as well, or something more traditional (Computer Science for example)?

To put things into perspective, when I was in school, I was in my early twenties and really had no idea what I wanted to do at that time. The technology field seemed really interesting though, and everybody seemed to be excited about the new developments there. I also realized just how bad accessibility for technologies at that time sucked, because I was having a terrible time just getting the materials I needed for school in an accessible manner. Other things were going on then, and it became necessary for me to find work instead of continuing my schooling. So I figured I could do both and learn more about the technology field and learn programming, hopefully to make things work a bit better.

So I bought outdated eBooks on markup languages, learned them, studied online resources and thousands of tutorials and blog posts, discovered how many were misguided, full of mistakes and just plain wrong when it came to using ATs such as screen readers with them, and had to eventually invent my own system to quantify accessible dynamic content management in a way that made sense and so that I could really make some progress in building interactive web controls accessibly and with consistent results. This also required that I learn visually oriented programming as well, such as CSS, in order to understand how everything fit together when combined as fully functional widgets. Since I can't see, I needed to find or invent tools that would allow me to do this, and luckily as time went on others in the field coming up against the same challenges were working on parallell projects and we were able to share ideas and make this happen.

None of these things were ever easy, and it's taken me over fifteen years to learn and build all that I have in order to help others so that they may not have to do the same as I.

If things had been different, I would have loved to get a CS degree. I don't think it would have made learning web development or ARIA any easier at that time, because these disciplins weren't around then as they exist now.

The value of the times that we are in now though, gives us a unique oportunity to provide real and accurate learning materials for all developers, because many of these technologies are sufficiently advanced and stable to allow for true education to be possible, where before they were not.

What upsets me about these circular conversations about ARIA, is the idea that we need to make explaining ARIA so simple that developers don't have to learn about accessibility.

I don't have any problem with making ARIA easy to learn, it's the idea that developers don't also need to understand how it effects accessibility that I find to be a disservice to future developers who actually might find these things interesting, instead of just onnorous.

Going back to the original topic, I would love it if technologies had in-built logic that automate accessibility especially when using ARIA. The fact is though that we don't have magically advanced systems such as these yet, and if future engineers are never trained in how to make them this way from the outset by having a firm grasp of the concepts involved, we never will.

I apologize for the rant, but I really having been banging my head against this particular wall for many years.

-----Original Message-----
From: _mallory [mailto:stommepoes@stommepoes.nl]
Sent: Monday, May 04, 2015 7:26 AM
To: Bryan Garaventa
Subject: Re: ARIA use in HTML other than for accessibility.

On Sun, May 03, 2015 at 04:18:22AM +0000, Bryan Garaventa wrote:
> > Yes tackling it in education is important. But also giving working developers resoruces to learn more without the expense involved in further education. 
> 
> I agree, but having gone through this particular gauntlet first hand, I also know that such educational resources must be first written by those who are most familiar with this knowledge in the field, which unfortunately does go back to education.
> 

Those whom I trust as a developer to tell me accurate things aren't teaching at university. They're working as engineers right now. Most of them are more familiar with the field because they're creating it themselves, at companies.

Last time I saw "web development" in higher education, it was still espousing <center> and <font> tags and everything in frames. In 2012.

You have gone through this before-- was it something most people in the field learned by themselves as well, or something more traditional (Computer Science for example)?

_mallory

Received on Tuesday, 5 May 2015 19:25:31 UTC