|
|
Log in / Subscribe / Register

Accessibility in Linux systems

Accessibility in Linux systems

Posted Oct 9, 2008 18:13 UTC (Thu) by riddochc (guest, #43)
Parent article: Accessibility in Linux systems

Disabilities come in many different forms. I suspect that my form is going to become increasingly common: I suffer from RSI, closely related but not identical to carpal tunnel, and thus largely untreatable.

I use a few tools to make my computing life easier:

  • Dasher, for text-entry. Unfortunately, being a university project, it's on a maintainership hiatus and most of the developers appear to have treated it as a means to a thesis rather than as a piece of software that needs stable releases for users that rely on it. The ability to copy all text automatically into the clipboard when movement stops was broken over a year ago, and hasn't been fixed. It takes more work to select all and copy by hand. And yes, they know about the bug - I told them.
  • A wacom tablet. Mine is an old graphire. Repetitive stress comes from repetitive actions - switching between different input methods helps to reduce that stress. I find that controlling Dasher with the stylus on the tablet involves the least physical movement possible for entering text, of the mechanisms that use physical movement at all. Unfortunately, the drivers for X can take a bit of twiddling to work. SaX on SuSE 11.0 can't produce a working xorg.conf file without help, and the newer versions of the wacom drivers don't play nice with the tablet or crash the X server at random. Yes, I've reported these bugs, too.
  • Workrave, a program to remind you to take typing breaks. The best way to keep RSI from happening, and to keep it from getting worse, is to take frequent breaks. Unfortunately, with my upgrade from SuSE 10.2 to 11.0, it crashes after running for a minute. I'm trying to figure out why, before filing a bug.
  • And now, the obligatory non-free software: Dragon NaturallySpeaking 10, running in Windows XP, running in VirtualBox. Through some miracle, VirtualBox can feed audio from also into XP and thus Dragon. The ~$200 version of Dragon can also take sound files from "digital recorders" in the form of wave files and transcribe them. It's nice, but it doesn't exactly integrate very well with Linux - it leads to a lot of temporary files and clipboard work.
  • Then there's Voxforge. If you want good speech recognition on Linux as much as I do, please read a book to them. The speech recognition engines (Sphinx, Julius, and HTK) are great, but what we need is a good collection of training data for speaker-independent recognition. This is tedious work, and needs volunteers. Please contribute!
  • I recently purchased an Orbitouch keyless keyboard. Unfortunately, they didn't put much thought into the arrangement of positions for letters - they're in sequential order. I'm working on remapping the keyboard so that going from one letter to the next will involve fewer physical movements. I haven't had the device long enough yet to really judge whether it's worth it.
As you can tell, there just aren't that many good options. People with disabilities face a really tough road trying to use computers - even more so if they intend to do any kind of software development. None of the options I've used so far are adequate for the kind of non-linear text editing necessary to do real work on code.

I've come to the conclusion that I have two choices: either stop using computers altogether, or focus my attention on improving accessibility myself. The truth is, the expectation that people will write code to scratch their own itch simply can't apply to people with RSI. The people who need these tools the most are the very same people who are least able to write that software, for obvious reasons.

Please, if you want to contribute and haven't yet, fix accessibility bugs!


to post comments

Accessibility in Linux systems

Posted Oct 9, 2008 22:56 UTC (Thu) by nix (subscriber, #2304) [Link] (3 responses)

OK, that keyboard is dramatically weirder than my Maltron.

(Regarding the key rearrangement: you might want to do as the Maltron
folks did and treat it as a large-search-space problem searching for
minimum finger motion across a large corpus of representative input: both
GP and simulated annealing seem likely to find something reasonably good.)

Accessibility in Linux systems

Posted Oct 9, 2008 23:06 UTC (Thu) by riddochc (guest, #43) [Link]

Yes, that was my plan. I'm working on code for a GP approach, similar to what has been tried for traditional keyboards. The approach is similar: I've collected letter bigram frequencies from text I've written (which I already use for training Dasher,) and I need to talk to someone in the nearby university's physiology department to get some help building a metric for the work involved in different arm movements.

Accessibility in Linux systems

Posted Oct 16, 2008 13:57 UTC (Thu) by nocomment (guest, #33767) [Link] (1 responses)

In the US Kinesis keyboards are similar and (somewhat) easier to find. See

http://www.kinesis-ergo.com/

Accessibility in Linux systems

Posted Oct 17, 2008 19:53 UTC (Fri) by nix (subscriber, #2304) [Link]

Indeed. I went for a Maltron mainly because they last (much) longer than
Kinesis keyboards do.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds