[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 17:06:41 -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 3:17 PM, Richard Stallman wrote:
We would like them to participate _usefully_.  Based on the
information you have provided, the only way to participate usefully in
the speech input area in the medium term is by helping replace

To recruit people for anything else would be make-work.

I'm sorry Richard. I am so sorry I have failed you in not clearly communicating the magnitude of the task in hand and what I was proposing.

The recognition system, in my opinion, is under 30% of the entire task. Building all of the user interface, tools, communications bridge is critical to the success of the speech recognition project. Only a very small portion, a replaceable portion of the task uses NaturallySpeaking. It is not enhancing NaturallySpeaking anymore than Emacs enhances the user's work environment. If you look at it from the editor environment, it can be replaced by VI, gnome edit, kate, spe, or any one of a dozen editors.

If we built the system, we would be 100% supporting the free software foundation speech recognition product. I'm sure some people were not part of the project could add a relatively minor component enabling the disabled user to increase the amount of free software available to them by a couple orders of magnitude.

Imagine that, a mostly free environment enabling disabled users to use an almost entirely free linux distribution. Without that support, choice is removed and they are forced to find nonfree solutions for all their common tools. I'm currently looking for a new editor. I've already picked my browser and e-mail client (Opera) and terminal emulator ( securecrt). not sure what I'm going to do about a word processor but buzzword, adobe's online editors looking pretty good. As I find more disabled user friendly tools, I will publish the list obviously elsewhere so that disabled users know what's useful to them.

This whole question becomes moot if there's another text input system that doesn't involve any more damage to my hands (or anyone else's) and is as fast and reliable as typing or speech recognition. I believe you implied there were some and I'm waiting to hear about them. I would love to give them a shot because I am also tired of being dependent on nonfree software. I am serious about that. I really am tired of it because I want to make it do things on my terms, not theirs.

Seriously, if there's anyway I can educate you as to the magnitude of the cliff you are getting ready to drive off of, let me know. I will spend the time, the energy, what ever it takes. I have seen two maybe three projects try to tackle what you are tackling now and they have failed miserably.

I am not just throwing stones. I want this to succeed but so far, you haven't laid out any sort of a plan indicating a path to success. This is not a project where you can just start cutting code and debug your way to a working application. It takes a lot of careful thought, careful planning from more disciplines than coding.

I've given up on the whole rights of the disabled versus free software discussion but strictly a technical, "make this project succeed" level, there needs to be much more conversation and planning. for example, what is going to be a strategy for dealing with different microphone acoustic profiles and various types of background noises (white noise, impulse etc.)? how are you going to do with the current inadequacies of the linux recognition environment? Are you going to deal with speech recognition on a machine distant from the user? Heck, I haven't even seen that solve for any of the blind support systems. They seem to assume that the machine the user is a machine the applications run on.

I will continue to support the UI side if somebody's willing to write code for me. I will give you my thinking and design efforts without any question but please, work out a plan, make it public with all the resources you need fully known. Reference existing recognition engines and show how you've got someone with the knowledge or someone willing to train themselves to get the knowledge necessary to handle those subsystems. Once that happens, I will have confidence that you are serious and this project has a chance at success and I know my efforts, my contributions won't be wasted.

reply via email to

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