the direction of speakup

Tony Baechler tony at
Thu May 2 04:26:36 EDT 2013

Hash: SHA256

This kind of approach described below won't work.  Back in the early days
of Window-Eyes and Windows 3.1 support, GW Micro did exactly what you're
describing.  Specifically, they designed the screen reader interface and
hired a development firm to code it.  It was very slow, crashed often and
generally didn't work very well.  I know from experience, still having the
3.5 inch disks for 1.0 and 1.1 before they did the complete rewrite
themselves.  WE 2.0 was a much better product and they learned their
lessons.  No, the Speakup and kernel communities should try to work
together somehow.

William suggested loading the user space daemon in an initramfs.  Well,
that's fine, but what if there is a kernel panic before the initramfs is
loaded?  I've had that situation before.  Also, what if a script in the
initramfs errors out and causes the next process (such as the Speakup
daemon) not to load?  I've had that happen before as well.  I was really
glad that I had Speakup at my Busybox shell so I could fix my boot
problems at the initramfs level.

On 5/1/2013 2:40 PM, Scott D. Henning wrote:
> Hi,
> I have been inactive here while studying cisco networking. I noted the 
> subject line and wanted to "listen" in. You bring up good points. ISA
> is gone today and support seems unneeded. Serial support seems as vital
> as monitor support for sighted users. When I reflect on how to bring
> useful code out for blind users, I imagine having users design the
> interface; detail the needs and then pass it off to a skilled
> programmer who could write it in their sleep. I realize this is counter
> to what you suggested here, but it makes sense to utilize the skills
> programmers have. They do not realize how the code they write affects a
> blind users understanding of what is presented on the screen. If they
> did, the code would come easily.
> I wonder if at times this is not a technical issue, but an ethical one.
> Do kernal writers know about the community of blind users? How hard
> would it really be for one of them to add this function? Intercept the
> text output to the screen and shove it out a port? I have often thought
> that if the design was known, what would it cost to have it written?
> this is a good thread. I am worried that speakup will wither if not
> kept at the fore of development. The world does not use text anymore
> and it is an invisible issue to most.
> Who would we work with to add speakup to a kernal, if your question to
> place it in user space comes back negative. That is it wont run early
> enough to voice all screen output and it must be in kernal.
> Thanks for bringing this up. Now I go back to the networking
> Scott

- -- 
Have a good day,
Tony Baechler
tony at
Version: GnuPG v1.4.11 (MingW32)
Comment: Using GnuPG with Thunderbird -


More information about the Speakup mailing list