Path: number1.nntp.dca.giganews.com!border1.nntp.dca.giganews.co
!nntp.giganews.com!newsfeed00.sul.t-online.de!t-online.de!newsfe
d.freenet.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From: Roberto Waltman 
Newsgroups: comp.arch.embedded,comp.os.linux.embedded,alt.os.lin
x.ubuntu,comp.os.linux.development.apps
Subject: Re: Speech recognition in lieu of lab assistant?
Date: Tue, 13 Feb 2007 18:44:03 -0500
Lines: 54
Message-ID: <20i4t21f9t4414i7n48m7jf01sc5rl5p2r@4ax.com>
References:  
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: individual.net eNO0hQ/XGPBsIvN98ofiNQ5wyScdedo/ofNt82TJ0A/FQ09Cv/
X-Newsreader: Forte Agent 3.2/32.830
Xref: number1.nntp.dca.giganews.com comp.arch.embedded:249445 co
p.os.linux.embedded:18509 alt.os.linux.ubuntu:10350 comp.os.linu
.development.apps:149533

>> Roberto Waltman wrote:
>> I wonder how difficult would be to add a speech recognition program ...

And "Mikael W. Bertelsen" wrote
>As I understand it, you write some test program yourself, and I assume you
>run them from a shell since you said "... while I type a few commands ...".
>My question is then, why not write a short shell script that will run the
>commands with an appropriate delay between them? You mentioned that it
>will only be a short list of options, which again point to using a shell
>script.


And "larwe" wrote:
>Sounds like a foot switch would solve all your problems.

Thanks but nope, (to both Mikael and Lewin.) When I wrote that post,
we were at the early stages of testing a few new boards still warm
from the reflow oven, and the choice of what test to run, and what
parameters to provide, was driven by the good or bad results from the
previous tests, going as far as writing new tests on the fly to
clarify some puzzling results.

Most of the tests were shell scripts driving simple program that
provide access to physical memory and CPU control registers from
userland Linux programs. The purpose was not only to verify
functionality, but also find errors in a new design. (And we did ...)

There was nothing deterministic enough that I could just insert delays
at predefined spots. And it goes without saying that the centerpiece
of the design was a BGA package that provided no convenient attachment
points for oscilloscope probes, so help was required.


And cs_posting@hotmail.com wrote:
>Learn to use the nose and big toe that the good lord gave you.
>Come on, everyone knows ctrl-alt-faceplant.
>And if you can't get your toe on the scope's single-shot aquire
>button, get yourself to yoga class.

That was always plan 'B' ... ;)


Things are more calm now, and I am moving into some automated testing
of a now known-good circuitry. I did not try the speech recognition
package that Jan recommended, but I'll do it soon to be ready for the
next time I need to use seven hands simultaneously.

Thanks for the suggestions,


Roberto Waltman

[ Please reply to the group,
  return address is invalid ]
--- Platinum Xpress/Win/WINServer v3.0pr5
 * Origin: Omicron Theta BBS (1:261/20)