From: Mike '/m' 

>Which should be about 10% of test variations...

Yup.  One very large phase of our QA test cycle for software is called
"negative testing".   That's where we provide non-valid scenarios
to the
software to see how it handles what it does not expect.

That phase was suggested by a 27 year old software engineer.  I guess I
must have hired one of the best and brightest that Microsoft missed....

 /m


On Thu, 17 Jul 2003 23:47:18 -0400, "Tony Ingenoso"
 wrote:

>Which should be about 10% of test variations...
>
>Writing code that works with correct inputs is no longer the challenge in
today's hostile environment.  One can think of processing error cases in
terms of smoothing on some bondo after the fact, or constructing an
imposing gauntlet up front.
>
>I sense some of the Bondo mentality still present at MS.
>  "Rich"  wrote in message news:3f171780{at}w3.nls.net...
>     It works correctly for all valid scenarios...

--- BBBS/NT v4.01 Flag-4