WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Web Analytics

for

From: Ryan Hemphill
Date: Apr 10, 2012 1:37PM

Hi Lucy, I understand what you are saying. I get it - I really do, but
technology-wise it's just not going to work in a Rich Internet App
environment. It isn't about what I want - I would much rather not have to
worry about the differences, it's a royal pain in the neck - but there are
only 3 screen readers that can truly handle Rich Internet apps at this time
(Focus Management and ARIA support) and the rest wouldn't be able to use
most those sites at all even if they wanted to. Granted I haven't worked
with ChromeVox much yet and that is a possible 4th.

It's annoying, it's upsetting and a whole list of other things that
shouldn't be said in public - but logistically, I can't see how we can move
forward and avoid these things. There is simply too much noise and most
dev/design teams would throw up their hands if they were dealing with a RIA
that needed to be made accessible, just like the guy at Google expressed.
This may be the reason they created their own SR. That being said, I have
been hammering away at that particular plugin and it doesn't work very
nicely in the RIA version of Gmail at all - or maybe I haven't figured out
how to use it just yet.

In most cases I would not be coding up tables or things like that any
differently - it's when we start getting into the interactive aspects.
Jumping back and forth through focus management techniques. Being able
to drop focus onto modal windows and use it in several different methods
without the screen reader going up in flames is a perfect example. I
actually had to take the AEGIS plugin (considered fully accessible, for
those who don't know) for modal/dialog and revamp it so that it would
behave like a typical modal window and work across multiple screen readers.
Part of fixing it required me to actually remove the role="dialog"
attribute because it seriously screwed up our ability to extend its usage.
ARIA actually REDUCED the accessibility of the widget. Imagine that! And
it didn't have the same effects on different SRs (NVDA and JAWS) until we
played around with it for some time. I can tell you some other bugs that
would have been introduced if we'd kept the native version, one of which
has the potential to break the whole site and kill the virtual buffer
permanently in JAWS, by the way. VoiceOver hasn't been handled yet, but
making that one work requires a different tweak to pull it off. Does it
matter? I would say yes. Purchasing textbooks cross-platform should
definitely be a priority use case for screen reader users. Would that
require compatibility hacks? Also yes.

This isn't something I love spending my time doing. It's something that
has to be done to insure accessibility of the site, which in this case is
for purchasing textbooks, by the way. SRs like Windows Eyes don't do focus
management at all, so we didn't even bother testing that SR because it
simply can't handle the interactive requirements of the site. Could we
make this site work for all SRs? Yes, we could as a basic page with no
bells and whistles at all. Could we make the site work for all SRs and also
have rich internet behaviors which involve focus management and possibly
ARIA attributes? NO WAY. Can I tell the marketing team to cease using
these features? They would laugh me right out of the room.

I go back to what I said before - we can't move forward if we're constantly
expecting to wait for the AT companies to get their act together and all
pull in the accessibility row boat in tandem. There is a good chance many
of the SRs aren't going to be considered usable for the web at all in a
short while, such as Windows Eyes, if they don't get their act together.

All that being said, I do appreciate your passion for the topic.

Ryan



On Tue, Apr 10, 2012 at 2:06 PM, Lucy Greco < <EMAIL REMOVED> >wrote:

> Hello:
> there would be no point in determining that a screen reader user is
> accessing your website because every single screen reader works differently
> than the other. You can't go out and detect every single competitive
> screen reader and expect to have a modified version of your website. So
> are you going to code your tables differently for JAWS then you would with
> chrome vox? Are you going to make your website change to voiceover
> compatible or are you going to work with NVDA or window eyes. Maybe the
> visitor is using guide or the full supernova suite.
>
> It's none of our business what the person is using to access our website I
> know there are people out there still using links but we don't code for
> that either. I know it's already been said on this thread but equal access
> means you made a good site that works for as many people as possible. Who
> cares what screen reader they're using. Who cares if they're using a
> screen reader the website should work and work for everyone. I don't want
> Amazon to know if I'm blind or not I don't want you to know if I'm blind or
> not I just want you to consider me a potential visitor to your website and
> make it work.
>
> Once all this is been said and done as a trainer I need to be able to test
> something at home on my machine and understand how it works then go to a
> client system and not have any surprises because they might be using an
> earlier version of the screen reader or the browser or the OS.
>
> Lucy Greco
> Assistive Technology Specialist
> Disabled Student's Program UC Berkeley
> (510) 643-7591
> http://attlc.berkeley.edu
> http://webaccess.berkeley.edu
>
>
> -----Original Message-----
> From: <EMAIL REMOVED> [mailto:
> <EMAIL REMOVED> ] On Behalf Of Tim Harshbarger
> Sent: Tuesday, April 10, 2012 10:38 AM
> To: WebAIM Discussion List
> Subject: Re: [WebAIM] Web Analytics
>
> Jared makes good points. I will try to make another good point.
>
> One of the big issues with this approach isn't technical but social.
> People with disabilities have experienced and do experience discrimination
> based on their disability. One thing people with disabilities seem to
> prize on the internet is having the opportunity to interact with other
> people and organizations without their disability becoming known. While it
> might be hard to believe in this day and age, interactions with
> organizations and other people can be qualitatively different if disability
> isn't one of the factors.
> Another factor in this attitude might be too many well intentioned people
> with little understanding. Basically, this approach is like saying "trust
> me, I know exactly what you need." I expect we all here are in a better
> position to understand what the user might need or want--but I expect
> everyone on this list with a disability has stories about well-intentioned
> people who felt they should make a decision on behalf of a person with a
> disability--and things ended up in disaster. But I am just guessing about
> this part.
>
> So, the population we want to design user interfaces for is going to be
> somewhat resistant to us trying to determine if they have a disability and
> then prescribing their user experience for them based on that information.
> I even suspect that most of them would be more comfortable answering a
> question like "Do you want white text on a black background?" rather than
> "Are you a person with a visual impairment?"--not to mention it probably
> will get more answers since not all people we might define as having a
> disability define themselves as having a disability.
>
> I do realize that when we design user interfaces, we have to make
> decisions based on what we think the users might need or want. I just
> think that people with disabilities would see a big difference between
> universal design and AT sniffing.
>
> That is just my two cents worth--or the denomination of your own choosing.
>
> Tim
>
>
> I definitely understand the desire of some of the people in this community
> to want that information about users--we all want to be able to give people
> the best user experience we possibly can design. With this approach, I
> think the problem we run into is people with disabilities have had a lot of
> bad experiences with that identification approach. Screen reader only
> pages, anyone?
>
>
>
> -----Original Message-----
> From: <EMAIL REMOVED> [mailto:
> <EMAIL REMOVED> ] On Behalf Of Ryan Hemphill
> Sent: Tuesday, April 10, 2012 10:10 AM
> To: WebAIM Discussion List
> Subject: Re: [WebAIM] Web Analytics
>
> When I read the reason for not wanted to be identified as using an
> accessibility technology, I gave a big sigh yesterday.
>
> You can't hide the fact that you are a screen reader user. Anyone with any
> understanding of screen reader behaviors would be able to detect your
> technology almost immediately. Some of them are very straightforward.
>
> (1) Clicking on a hidden link/anchor/button within a page. While this is
> listed as an accessibility technique/practice, it's a red flag that you are
> a screen reader user. No one else would see it in the first place and
> there is definitely no one that would be able to click on it.
>
> (2) Tabbing a lot. One focus jump after another going all over the entire
> page (let along the entire site) unless it was a form would immediately
> give a high probability of the user leveraging a screen reader or some
> other accessibility software.
>
> (3) Tabbing + Click event. This one would be a very strong indicator. If
> the user was tabbing a lot and then clicked (via the simulated click event
> that screen readers all use), the probability that the user is navigating
> via screen reader is very very high. Why else would anyone tab through an
> entire site only to use a mouse click once they have found what they were
> looking for?
>
> (4) I can detect, at this time, whether you are using JAWS, NVDA or
> VoiceOver in a heartbeat. I have 2 methods I could use in JAWS, and might
> even be able to figure out the version of JAWS with a little bit of extra
> research. Given the information you have just provided (thank you for
> that), I will avoid releasing how I can do this, but suffice to say - it
> isn't hard once you really understand the behavioral differences from each
> other and keyboard/non-screen reader users as a whole.
>
> We need to revisit this issue. There is just no way you are going to be
> able to hide many accessibility technologies from any developer on the web.
> If someone has nefarious purpose for detection of your screen reader
> technology or wants to detect the likeliness that you are a keyboard only
> user that might have motor skill impairments, for example, you are not
> going to be able to stop them.
>
> Furthermore, you are creating a situation that makes it extraordinarily
> difficult to deal with the differences between screen readers. The
> compatibility issues that present themselves in rich internet applications
> already make it so that JAWS, NVDA and VoiceOver are the only possible
> means to handle things like focus management because it is that complex to
> begin with. Add on the fact that they each handle it differently and you
> are really looking at a major break between the desired avoidance of
> detection and reality of making Rich Internet Apps work for these programs.
> Developers who are doing Rich Internet Apps need to know what you are
> using especially when the RIA becomes complex.
>
> In our current situation at my company, we are required by law to provide
> accessible solutions for screen reader users, but there is more than one
> Rich Internet App in our company that require knowledge of which (JAWS,
> NVDA or VoiceOver) screen reader is being used to insure failures
> (interaction or formatting) don't occur. There is too much going on in
> these apps for the SR software to handle it without some help as well. I
> could go on for at least an hour or two about how these situations come
> about and what we are doing to resolve them but mark my words, there is no
> question at all that we need to know what we're working with.
>
> I get why no one wants to be identified. No one wants their information
> broadcasted to companies that want to exploit their data. It is invasive
> and unsetting - I completely understand the objection.
>
> But it isn't going to allow screen reader technologies (or others for that
> matter) to maintain the pace of development that we are all witnessing even
> now. I even remember seeing a post about half a year ago by a Google
> employee that was stating how making some technologies accessible at this
> time was extraordinarily difficult. This is Google we are talking about, a
> company that hires super-geniuses to write their software - and they're
> saying they can't the target. Now granted, I'm sure that there are those
> among you that see this as an excuse or lack of knowledge on that person's
> part - but for a massive technology driven company to have an employee
> state publicly their doubts in creation of RIA accessibility - that's a
> very bad sign. And we aren't doing them or any other design dev teams any
> favors (or the user base it affects) by hiding the fact that you are a
> screen reader user or have some other a11y tech-specific need.
>
> If this post needs to move to a new discussion (probably not a bad idea) I
> would like to continue this conversation further with anyone that wishes to
> state their opinion. I am open to harsh criticism, so fire away, but I
> feel very strongly that this perspective is going to hold back
> accessible-friendly technology for the web in a critical and unfortunate
> way. Don't get me wrong, I care very much about doing the right thing, but
> as the assessment of a 15 year veteran in design/development, there is no
> way this is going to work in the long run - and I am not the only person
> that would tell you that.
>
> We need to be open to telling the dev teams what tech we're shooting for.
> A no-holds-barred approach to keep up with the pace we are seeing. It's
> not going to slow down and you can't predict what is coming up next. I am
> very unsettled and concerned by this issue.
>
>
> Ryan.
> > > > > > >
>
> > > >



--



Shipping is a Feature...Perhaps the Most Important Feature.