Subject: | |
From: | |
Reply To: | |
Date: | Sat, 19 Dec 1998 21:44:42 -0500 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
At 4:15 AM -0500 12/19/98, Jerry van Dijk wrote:
>> have not seen any data proving that the CMU Capability Maturity Model levels
>> have anything to do with quality or cost. Admittedly the level of an
>> organization appears to be correlated with their capacity to estimate cost
>> and delivery time. However, usability and maintainability are also very
>> important.
>
>Without a process, there is no way to influence the quality level. With a
>well defined and well embedded process you will not make the same mistake
>twice. Therefor the CMM process influences both cost and quality.
>
>Jerry.
Many statements on the surface appear to be true. A couple of hundred
years of recent science suggests we better do some experiments before
we assume something to be true. In the previous paragraph, there are
three statements; call them A, B, and C. Your logic is:
A and B => C
If A and B are true, C might partially follow. Many people do believe
both A and B and are willing to spend a great deal of money based on
this logic. However, statement A is not experimentally provable.
Unless you define everything that influences quality a process.
Then it is not even an interesting statement.
Statement B's truth implies there exists processes which can be
measurably "defined" and "embedded" to the point of not repeating
a mistake. True only if all mechanisms in the process are 100%
reliable and there are no "bugs" in the process. Allowing for a
less rigid interpretation of the statement, the truth would then
follow if experiements showed a reduction in "repeated" errors.
And finally, statement C's implication about cost does
not even logically follow.
I did not author the first paragraph in this e-mail. However, I am
also looking for the experimental evidence which correllates CMM
level with low cost and high quality. Then there will need to be
further research to indicate cause.
Doug Smith
[log in to unmask]
|
|
|