Software Engineering - Well-Differentiated Quality Models

Author

Well-Differentiated Quality Models

Eddie Burris

2004-07-24, 3:56 am

I have been reading articles on software quality models (Boehm,
McCall, ISO 9126) and noticed that they all strive for a set of high
level quality characteristics that don't overlap. For example, none
would suggest having both testability and verifiability as quality
characteristics because there is too much overlap between these two
concepts. I don't understand why a well-differentiated quality model
is that much better than one with overlapping characteristics. What
practical problems can result from a quality model with overlapping
characteristics?

Eddie Burris

David Lightstone

2004-07-28, 9:04 pm

"Eddie Burris" <burrise@umkc.edu> wrote in message
news:6efa7e30.0407231738.3d97ed69@posting.google.com...
> I have been reading articles on software quality models (Boehm,
> McCall, ISO 9126) and noticed that they all strive for a set of high
> level quality characteristics that don't overlap. For example, none
> would suggest having both testability and verifiability as quality
> characteristics because there is too much overlap between these two
> concepts. I don't understand why a well-differentiated quality model
> is that much better than one with overlapping characteristics. What
> practical problems can result from a quality model with overlapping
> characteristics?
>
> Eddie Burris

Its the difference between a linear model for a spring and a non-linear
model for a spring.
One is more accurate and realistic. The other allows you to reason about the
problem.
Needing to solve a problem the authors favor making a guess based on some
reasoning rather than no reasoning

The conclusions based upon the linear mode are not always valid
(ie don't extrapolate to far)

Sponsored Links