Re: [Accessibility] Call to Arms

[Eric S. Johansson]

 On 7/28/2010 5:28 AM, Jason White wrote:
> Willem van der Walt<address@hidden>  wrote:
> > I think that those of us who are serious about a11y on none-proprietary
> > platforms and also understand the four freedoms as stated by the FSF will
> > end up working together anyway.
> I agree, and I welcome the FSF's decision to take a leadership role in this
> area.
>
> I support the development of free (as in freedom) accessibility-related
> software for free operating systems, as well as improvements in the
> accessibility of free programs. I think this essentially captures the ambit of
> this forum; it is concerned with the intersection of free software and
> accessibility for people with disabilities. It is not concerned with
> proprietary software or the development of libraries and tools that
> are designed to interoperate with it.

1) free software also affects the ability for disabled people to be part of our 
modern society

if you prevent them using a nonfree component through implicit or explicit 
actions when there is no free replacement, then you effectively deny them 
accessibility tools.

Another analogy I came up with last night might illustrate the problem. I'm not 
sure it works but I think it fits.

If a doctor has a multidrug solution that cures his patients but one essential 
drug is encumbered is application by drug company, is the Dr. right to deny his 
patients relief until the encumbered component has been freed?  how ethical is 
the Dr.'s stand if the delay is a short time versus a long time?

Yes, I see the issue of making an accessible *system* as critical as providing 
medical relief.

2) free components and a free system

 I believe the description of what I and a few others want to do, is make a 
collection of library and tools that will be central to the solution of the 
totally free environment. Unfortunately, we cannot test them until the entire 
system exists. We can't do usability tests until the entire system exists. Yeah, 
broken record I know.  If we used the totally free dragonfly library as the 
base, we could make progress in parallel or even in advance of the free software 
recognizer and then plug-in the new recognizer when it's done.

3) project scope

Even though the Sphinx systems are inadequate, using them and the size of Julius 
or Simon can give you a feel for the scope of what you're trying to do.

it out every once in a while when you're ranting about things because of what 
you've learned, you stumble across something that says you might have been wrong 
but you're not sure because you haven't had time to research it.

http://www.bltek.com/virtual-teacher-side-menu/sonic.html

At the very least, this sonic research paper given below give you really good 
feel for the scope of what you're trying to do.  at the very best, it's a core 
for starting the project. The license is probably not acceptable, at first 
glance it looked like a dual license with a BSD-ish slant. In any case, it looks 
like the project is effectively abandon ware so maybe there's a chance of 
relicensing. I have no idea where it sits Technically in comparison to Julius 
except development seems to be continuing with Julius


a couple of interesting papers that apply. It's really weird. Most of the papers 
on large vocabulary continuous speech recognition seems to stop roughly circa 
2004. It's like the community decided en masse that the problem couldn't be 
solved to their satisfaction with what we know today and there is no money for 
deep research to fix the knowledge problem.

www.icsi.berkeley.edu/~dpwe/research/talks/TUT-ASR-1999sep.pdf

tanaka-www.cs.titech.ac.jp/publication/archive/12.pdf
http://www.era.lib.ed.ac.uk/handle/1842/1185

--- eric