[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Accessibility] Call to Arms

From: Eric S. Johansson
Subject: Re: [Accessibility] Call to Arms
Date: Wed, 28 Jul 2010 13:40:36 -0400
User-agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv: Gecko/20100713 Thunderbird/3.1.1

 On 7/28/2010 1:21 PM, Christian Hofstader wrote:
Eric, et al, Sorry for being so presumptious as to suggest an idea for doing something I thought might be helpful. I'm not an expert in this stuff - not even close - but, as there is so much interest in programming by voice and more general speech recognition tasks, I thought I might toss out a few ideas that came to mind. So, don't bang your head over me as I'm not terribly qualified to do more than dance in the dark,

I'm sorry, I do owe you an apology. I'm afraid it's from my fatigue caused by going around this barn so many times as new geeks come in who have little experience a speech recognition and suggest the same things over and over again. sorry for being a jerk.

I am torn between pointing out with the edge of the cliff is, sitting back and watching people dance on the edge of the cliff knowing full well where they will end up in about 1 to 2 years or, getting behind them and pushing. In either case, it's a lose proposition because foreknowledge is frequently seen as negative because by gum, their enthusiasm and energy will overcome experience and wisdom any day. ;-)

There are more options for programming by voice than the one I have outlined so far. I didn't want to overwhelm people. The problem is we seem to have an intractable barrier. The only speech recognition system we can use in anger is NaturallySpeaking and building speech recognition applications that even vaguely use NaturallySpeaking is in conflict with the free software foundation philosophy.

I'm have a really hard time seeing how you are going to make any progress if you don't have a real recognition system. If you build something on Sphinx, you have a small-scale system that can't be used for any serious code writing. If we cut people's vocabularies down to a couple thousand, we might be able to do that but unique symbol creation will be even more difficult. There's no guarantee that it will scale in terms of what people want to do and it might backfire and drive people into the arms of NaturallySpeaking because of using it where it really doesn't belong.

This feel sort of like the way nonfree services were using free software on the cloud.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]