Greg Woodhouse wrote:
> It seems to me that there are two threads of discussion here that are
> not at all merging. One issue is whether testing can and should be made
> cheaper. Maybe it can, but testing is the last line of defense in
> software quality, and is highly problematic, relying essentially on
> chance to hit upon the right combination of conditions to cause a
> defect in the software to manifest itself. It may be that extensive

Hmmmm.. I see what you mean, but I believe it's possible to use the 
testing and certification process as the _first_ line of defense in a 
good quality control program.  The actual certification from say, CCHIT 
would occur last, but high marks on certification should indicate that 
testing, overall was an integral part of development.

> testing (in the form of clinical trials) is the state of the art in
> medicine, too. But in software we can and should do better. Every
> computer science student learns in his or her first semester that the
> halting problem is insoluble, and so we all throw up our hands and say:

Gregory is right on this count. One of the first things I learned in 
computer science school was "If houses were built like software, they 
would all fall down."  I've always tried to do better than that in my 

Richard Schilling

Yahoo! Groups Links

<*> To visit your group on the web, go to:

<*> To unsubscribe from this group, send an email to:

<*> Your use of Yahoo! Groups is subject to:

Reply via email to