ideological speed bumps

Eric S. Johansson esj at harvee.org
Mon May 17 14:04:00 BST 2010


On 5/16/2010 11:07 AM, Hugh Sasse wrote:
> One would have hoped that 19 years after the Americans with
> Disabilities Act, and 15 years after similar UK legislation was
...
> (unfortunately commercial) system" problem immediately.  But a
> standard could drive innovation in some cases.  This hasn't worked
> perfectly for HTML and browsers, but it has worked to some extent, I
> think.

warning: I had way too much caffeine yesterday and I'm running on about two 
hours sleep so this may not make sense. I would delay responding normally but, 
I'm running short on time in the near future and don't want to lose the thread 
hence the too much caffeine and two hours sleep.

Hugh, I think this is a good segue from some of the other things I was talking 
about. As one person said (effectively) wwUd, what would Ubuntu do? Canonical 
has a long history of crossing the boundaries between commercial and open source 
components and so they would be a natural to help support an effort for handicap 
accessibility using both commercial and open-source components.

I also have an unshakable belief that across machine solution is the best one 
for the short and long-term. As of said elsewhere, I will be a broken record and 
say here that accessibility features should be owned by the disabled person. If 
you're blind, you own the text-to-speech engine, if your hands are broken, you 
on the speech recognition engine. If all you can do is use a unicorn stick, then 
you on the stick and the special keyboard. It is not the responsibility of the 
host running the application to provide any accessibility tools or capabilities. 
It is the responsibility of the application host to provide you access to every 
single function the application provides publicly to any interface. This also 
includes the information the application operates on.

I prefer this solution because it is far more testable with lower resources on 
any application developer therefore they have no excuse for not making it work. 
No more will we hear the "we don't have a disabled people therefore we can't 
test text-to-speech or speech he keys or or or". A standard API drivenn test is 
simply and easily validate accessibility functionality.  one great test for the 
usefulness of the API in an application is if you can extract the GUI and make 
it run separately from the application. Like I said, all interfaces, one API.

Another reason why across machine solutions are better is the simple issue of 
licensing. I'm not on the run NaturallySpeaking on every single platform I use. 
Never mind most of them are Lennox but I would need to run on something like 30 
to 50 machines per year and that $500 per license, that would break the bank. 
But if I have a single license on my portable platform and connect to virtually 
any machine, this is a good thing because it reduces the number of instances of 
closed source software you need to use and maximizes OSS visibility for disabled 
users and hopefully their employers once they can now do more things with a 
linux computer then they can Windows.

What this interface looks like is, quite frankly not clear. For example, one of 
my favorite tools to build would be the enhanced dictation box. How do you 
define that? Is it an application built on top of a standard disability API? 
Don't know.

on the legislative/legal question, one axiom I've come to embrace is law cannot 
prevent what technology enables. There lots of practical examples of this in a 
disability world, law cannot prevent disabled people  from being completely 
locked out by application technology. Every time a new revision comes, a new 
technology arrives (think iPhone), or a new business practice (cloud computing) 
arrives on the scene, we keep fighting the same old battle over and over again 
about putting in disability access and business criminals whingeing about how 
hard it is. If we can fight back with our own technology which reduces cost of 
inclusion, and, if it's built in, make the cost almost nothing, then then 
legislative efforts have a chance at succeeding. This is especially true if we 
could use public watchdogs to run test suites to say "fails or passes". If they 
want the answers of how, then they pay the watchdog organization for that 
information so the watchdog organization can keep doing what it does.

Anyway, got to go do other things which will hopefully involve sleep sometime 
the next few hours. :-)



More information about the Ubuntu-accessibility mailing list