Re: User control (was Re: Non-underlined Links)

by "J Bryan Kramer" <jbryankramer(at)email.msn.com>

 Date:  Sun, 23 May 1999 19:59:59 -0400
 To:  "wistfulee" <wistfulee(at)shaka.com>,
<hwg-theory(at)hwg.org>,
"David Meadows" <david(at)heroes.force9.co.uk>,
"Harold A. Driscoll" <harold(at)driscoll.chi.il.us>
 References:  oemcomputer
  todo: View Thread, Original

I'm a beta tester for MS, so let me shed some light on this. First all the
beta's I've been on last for months and often over a year. I've betaed
Win95, OSR2, Win98, Win98 SE, Win NT4, Win NT 5 and Office 2000 along with
other minor stuff for MS. MS also has an in-house usability and testing lab
and they bring in groups from nearby areas to test the software in the lab.
They supposedly have hundreds of different hardware setups and they vary the
groups from new users to power users.

However they will never get all the bugs out for several reasons. The first
big one is this, these software packages are so huge, millions of lines of
code and hundreds of different modules interacting together that it is
mathematically impossible to remove all the bugs. I really mean that this
has been proven mathematically using number theory or what ever.

Secondly, in the Intel world, there are so many different system components
available: 50 different kinds of video cards, a hundred system boards, 75
different modems and so on that all the possible system configurations
cannot be tested. There are millions of possible combinations. So all this
hardware mismash leads to unexpected conflicts.

Finally you have to throw in bad programming practices on the part of all
the other software houses. The major cause of crashes is that someone
overwrote memory that belonged to the operating system or another program.
Microsoft has extensive training and resources available for developers and
also a software certification program to try to help in this area.

So combining the size of the software, unknown hardware combinations and bad
programming by other software houses you can see that its no wonder that we
have problems. At least MS does post frequent patches, generally free, as
problems pop up. Some other software houses just don't do this.

Other companies also cut the beta process short and are even more prone to
problems. Corel was notorious for this in the past, I don't know if they've
corrected the problems recently.

J Bryan Kramer
jbryankramer(at)email.msn.com
ClubWin Team 9


----- Original Message -----
> I have a question solely from a user's standpoint: why isn't software
> passing a usability test before distribution??  I am tired of buying &/or
> using (especially buying)software that doesn't work, has bugs (I guess the
> definition of bugs is that it doesn't work<S>), is incompatible with other
> things I may have on the puter & I have to install fix after fix
(disguised
> as "updates") to get it to work.  M$ is most famous for that, but they
> aren't the only ones.  It seems to me that what I am trying to learn here
is
> hampered by every company having different ideas of what is important, not
> every company trying to accommodate what users want to do.  Is it that by
> trying to make things easy for the people that follow,

HWG hwg-theory mailing list archives, maintained by Webmasters @ IWA