[Access-activists] Re: [Accessibility] Call to Arms

[Eric S. Johansson]

 On 7/27/2010 11:43 PM, Bill Cox wrote:
> On Tue, Jul 27, 2010 at 6:24 PM, Eric S. Johansson<address@hidden>  wrote:
> >   On 7/27/2010 8:57 PM, Bill Cox wrote:
> > I thought the discussion was worth having
> > because this is a free software foundation sponsored list and if they say,
> > thou shalt do only free software, then that is their right and we should
> > respect that.
> This gets to the heart of my greatest misgiving in having FSF leading
> FOSS accessibility development.  I think that FOSS is pretty much the
> only way to get programmers with disabilities from around the world to
> volunteer to work together to write this stuff, so I see significant
> synergy between FSF goals and the goals of the volunteers.

I think the best way to do it is with a Chinese wall type partition. We have 
this list for the majority of the work, supporting only free software and then 
we have a separate list for dealing with a transitional stuff. We also have 
separate repositories etc. It will be kind of interesting because the only way 
we can do testing in the beginning is with nonfree software.

> However, will the volunteers see it that way?  Will programmers who
> use Windows and Naturally Speaking, or Windows and JAWs hesitate to
> join a FSF community?  FSF has broad reach, resources for servers, web
> development, and a well recognised brand, in addition to Chris working
> full time.  So, it seems like a great foundation to organise things
> from.  Will the community converge here?  Would we be better off
> building an on-line community of volunteers dedicated to
> accessibility, and only accessibility?  Can that be done through FSF?

Who the heck knows. The ideals put forth by the free software foundation are 
worth living by. The compassion conflict bothers me. Most people I know who are 
doing open source work with speech recognition tools are pragmatic and accept 
the fact that NaturallySpeaking or Windows speech recognition are the only 
recognizers worth using. They are treating it as a "solved problem" because it 
really f-n hard to replace.  this is one reason why I suggest both virtual 
machine and wine solutions as encapsulation.

one source of volunteers we will not see are former Dragon or nuance employees. 
The employee contract has a significant punitive aspect to any employee working 
in the same field. I believe the noncompete timeframe something like five years. 
And we have the problem of "inevitable disclosure" which is another trick former 
employers use to retaliate against employees going to a competitor. They have 
successfully argued that just by knowing copy secrets, a former employee will 
disclose by virtue of simply being there and working on the same problems. No 
sir, we don't need those problems.

Developing the recognizer et al. may need one or more graduate departments with 
boards of grad students. If memory serves, expect to spend something like $50 
million developer bux.

But I'm not going to depress myself any further with these thoughts. When I have 
the next few spare moments, I'm going to do some more brain dumps on adapting 
applications to speech recognition, mismatches between GUI and speech user 
interfaces, programming by voice, incremental steps

--- eric


> Bill