[nfbcs] My Opinion and Experiences with Accessibility

[Nicole Torcolini]

This email is in response to some of the threads that have been going on
over the last two or three days. I have separated it out, though, because I
hope that people will read it in it's own light instead of in the direct
shadow of some of the other messages.
This message might get a little lengthy, and I might have to finish it
another time. The following are my opinions, which are based off of my
experiences. Feel free to dispute them, but I ask that you please do it in a
polite manner. So, here goes...
Oh, and one other thing. All of this is in reference to web accessibility,
since that is the area with which I am most accessible.
 
It may be hard to define accessibility, but I think that it is possible to
have at least some sort of definition. To start, accessibility means that it
can be used with a screen reader. Okay, so what does "can be used with a
screen reader" mean? This might not be all of it, but, as a start for
defining  guide lines for developers, it is better than nothing:
1. Everything that can be achieved with the mouse can be achieved with only
the keyboard
2. Focus is handled correctly
3. ARIA roles and attributes are used and maintained correctly
4. HTML elements are used correctly
This might not seem like a lot, but these four areas have a lot of content.
For example, JMHO, number 3 covers issues like notifying the screen reader
when an area of the web page changes using aria-live. Number 4 could include
things like using headings when appropriate and not using tables purely for
layout purposes with stuff that is not data.
 
I don't think that developers deliberately decide to make their websites
inaccessible. A lot of developers don't know how to make their websites
accessible. Not everyone knows about ARIA, and, even if they do, there is
not a good way to verify that all of the ARIA stuff is correct without using
a screen reader. Even if a screen reader is used to test a website, most
sighted developers don't have the screen reader proficiency that we do, and,
even if they do, they might only have access to certain screen readers, so
there is no way to verify that it works with all screen readers. I can't
tell you how often I test something with screen reader x, and the developer
says "I don't understand. I tried it with y or z, and it worked." Usually,
once I explain, they understand, but it is extremely frustrating for them
when they put in the time and effort to test, but find out that it does not
work with a certain screen reader. To that end, some of the accessibility
problems are because screen readers act different and implement the ARIA
spec differently. I'm not suggesting that all screen readers should be the
same, but there are certain things that would make things easier if the were
consistent across screen readers. I also often run into the case where the
developer has tried, having good intensions, but has misunderstood what
certain ARIA attributes do. JMHO, the ARIA spec needs to give more
information.
Accessibility might not be the absolute first thing on the list when a
product is created, but it needs to be on there fairly early, long before
dogfood or even fish food. Accessibility is not something that can be added
later. The analogy that I like to use is adding ramps and elevators to a
building after it is built. It just does not work, or not that well. Even
though accessibility may not affect that many people, it needs to be treated
the same way that security is treated; it needs to be a "launch blocker".
Okay, so maybe it is not realistic to expect a product to be absolutely
perfect in terms of accessibility before it sees the light of day. Then how
do you determine what is enough? You make levels of accessibility, starting
simply with item number one on the list above. No, it's often not enough for
good screen reader support, and it usually keeps people from coding
themselves into a corner. There could be different levels of accessibility.
If a product is large, then different levels could be assigned to different
parts of the product based on how often that part is used.
In order for a product to be accessible, it sometimes takes the entire
company being committed to accessibility. Sometimes, developers want to make
the product accessibility, but they think or have been told to focus their
efforts in other areas. So, often, the best approach is a top down approach.
If someone high enough in the chain can be convinced that accessibility is
important, then he/she will tell the people whom they manage to work more on
accessibility.
Unfortunately, there is no quick and dirty guide to making a website
accessibility. If you want to make it really accessible the rather easy way,
then just use native HTML, but, in today's world, that just does not scale.
Developers need to understand ARIA roles and attributes, keyboard handling,
and focus handling. They also need to understand what screen readers cannot
do, such as detect most CSS or recognize images.
Finally, yes, often, the best way to test a product for accessibility is to
have users with disabilities try it. If a company is large enough, there may
be a fair number of employees with disabilities. Also, there are sometimes
employees who do not have disabilities but who work on accessibility who
know how to use assistive technology, but these are probably not as
prevalent in smaller companies.