[jawsscripts] Petition
- From: "Jim Snowbarger" <Snowman@xxxxxxxxxxxxxxxx>
- To: <jawsscripts@xxxxxxxxxxxxx>
- Date: Sat, 6 Oct 2012 00:55:59 -0500
In case you have not already seen this, a link to a grass roots petition to FS
to improve product quality:
http://www.ipetitions.com/petition/freedom-scientific-bug-fix-problems
Ok, I'm gonna wax eloquent for a bit.
I signed this petition, not because I don't have respect for the guys at FS and
what they have accomplished. Nor because I don't have sympathy for the
enormous challenge they face, trying to keep up with all the changes that
continue to prolipherate. Quite to the contrary, I do respect their
accomplishment, and think they've got a tough job. , For 15 years, I have, and
continue to, benefit from their work. I remain committed to JAWS, because it
is still the most powerful tool I have available to do my work.
I say, if NVDA can do it, and JAWS presently can't,
just give me a few minutes, and it will.
But, I signed because I generally concur with the basic sentiment expressed in
the petition that JAWS is a critical piece of software for me to conduct my
life. It is not a game, or an incidental amusement, it's a critical tool.
And, as such, it deserves some serious quality control attention.
To be fare, it's not just JAWS that has the bugs. Buggy software is the norm
these days, everywhere you look. Most of it has some imperfection or other,
and much of it is downright experimental, shouldn't have even been let out of
the lab. Much of this is driven by the pressures of a foney ephemeral market
place that doesn't know what it wants until it sees it, and then it likes it
for a day and gets bored, Move on. Just throw that away. Who cares, it
probably had bugs in it anyway. The problem is so pervasive in the industry
at large that, people like me, who write safety-related software are forced
through unbelievable rigor to prove, six ways to Sunday, that everything works,
exactly as it was intended to work, every single time. No room for mishaps.
It doesn't crash, you never need to reboot it. It doesn't sometimes do this,
and sometimes do that, it is dependable.
For every module, or logically related group of modules that is developed, a
rigorous and exhaustive set of unit tests are defined and implemented, to run
that unit through all it's paces, even to the extreme of having to define
adequate test cases to cause every single line of code to be executed. And,
before the entire system is built, all modules are run through their unit
tests, and the results confirmed. Then, the entire thing is assembled, and a
whole batch of integration tests is run on the assembled system. Those tests
accumulate as the product evolves. And, with every new release, old tests are
run again to prove that stuff that used to work, still does, exactly as it did
when it was first introduced, as modified in subsequent releases.
They make us go through all that pain and suffering precisely because the state
of quality in the industry at large is so bad. Developers seem to just kind
of get it working, and then throw it out there for the consumer to "enjoy".
Notice the quotes.
But JAWS isn't a foney, funny, throw away program like that. It's an important
tool.But, when you think about it, doing all the testing such as what I
described for all of JAWS, on all the operating systems, and service packs,with
different application versions, video cards, , and I don't know, processor
manufacturers? Surely not. But, anyway, that test effort is HUGE. I'm sure
they do testing. I'm just not sure what form that takes.
Anyway, if we get them to be more rigorous, it is going to take them longer to
do stuff.
The rate of releases would slow down. That means the tick rate on your SMA
slows down too, and your SMA lasts longer.
And, FS's income would fall, even as it works harder.
Problem is, the monkeys have to be fed.
So, FS earns a fixed amount of income by ticking the SMA's and a fixed rate.
If we compel them to release fewer features, but with higher quality on the
one's they do release, they can get their tick rate, we can get our quality,
and we can know that we paid a price for that quality, only if we believe that
the number of features we used to get per major release should be considered
the norm, and so dropping back from that so-called norm, is perceived as a
cost. But, if the so-called norm was actually artificial, in that the cost of
persisting bugs, and the aggravation of dealing with new features that don't
work quite right, made that artificial norm less than optimal, then the
relationship between that, and the situation we were pondering of fewer
features but higher quality, is irrelevant,because we are not supposed to get
as many features as we are now getting. If we were, it would be possible to
provide them with the quality level we expect. It apparently isn't, so we're
not.
But, we were getting that many, which means stuff was not like stuff was
spoasta be.
But, that should be abnormal.
In other words, the higher quality, fewer features case should be the true
norm. Can you look at simply moving from fantasizing an artificial norm, to
mentally adopting the true norm as a cost?
What do you know, higher quality, zero cost.
So anyway, I'm sure I'm just preaching to the quire to say that JAWS has, for
ages and ages, been plagued with a disturbing degree of variability, as well as
regression as new releases are poked out. Perhaps , it would make more sense
to release once certain quality objectives for the required feature set have
been achieved, rather than because it is October again.
I have always lived by the aphorism, keep your older versions handy, for a very
good reason. And so, I do.
Anyway, because this is such an important tool for us, It seems good to
collectively remind FS that, while we thank them for their work, we'd like them
to do a little better job focusing on achieving, and maintaining product
quality.
Ya think so?
__________�
View the list's information and change your settings at
http://www.freelists.org/list/jawsscripts
Other related posts: