To: Paricia K. Lawlis, Ph.D. et al.
From: Bob Leif, Ph.D.
Your paper concluded, "We observed improved cost and schedule performance
with increasing process maturity. Specifically, the least mature
organizations were likely to have difficulty adhering to cost and schedule
I agree that your data is suggestive that CMM level correlated with
predicting cost and schedule. However, the decrease in the scatter of the
data with CMM level indicates that CMM improves precision. Organizations
following CMM appear to be able to more accurately predict schedule and
cost. Unfortunately, our discussion concerned accuracy. Do they produce a
better product; rather than do they keep the administrators happy? In fact
your data could be interpreted that CMM provides extremely useful data to
the contractors. It tells them to increase their scheduled time and cost to
realistic values. I have no argument with this. However, increased precision
(repeatability) can lead to worse performance. One consistently misses the
I will admit that data mining, with all of its difficulties, is about all
that can presently be done, especially with minimal funds. I hope that some
studies will be done correlating the costs of the Year 2000 "Bug" with:
programming language, operating system, software development methodology,
and training of both software engineers and managers. The last item is of
importance because the Year 2000 problem is NOT the result of a software
bug; it is the result of poor management. Firstly, the existence of Year
2000 bugs which require significant costs to repair are very often a symptom
of the state of the software. In the case of DoD, year 2000 problems which
are the result of ignoring the Ada mandate should result in the person(s) or
organization(s)responsible for the project being held accountable.
> -----Original Message-----
> From: Team Ada: Ada Advocacy Issues (83 & 95)
> [mailto:[log in to unmask]]On Behalf Of Patricia K Lawlis
> Sent: Monday, December 21, 1998 11:29 AM
> To: [log in to unmask]
> Subject: Re: Language Efficiency was RE: Choose Ada flyer
> >> also looking for the experimental evidence which correllates CMM
> >> level with low cost and high quality.
> Perhaps my previous message didn't get through, or perhaps it was
> misunderstood. The paper I previously referenced, "A
> Correlational Study of
> the CMM and Software Development Performance", published in the
> Crosstalk in
> Sept of 95 and available at
> reports on some very nice experimental research done for a
> Masters thesis. It
> does, in fact, show statistical correlation between CMM level and
> cost (also
> Indeed, much more research is desirable. It is not easy to get
> good data that
> stands up to controlled research conditions, but at least some
> good data is
> out there. There should be a lot more available by now.
> Patricia K. Lawlis, Ph.D.
> c.j. kemp systems, inc.
> P.O. Box 586
> Fairborn, OH 45324-0586
> Fairborn office: (937) 878-3303 (voice and fax)
> Phoenix office: (602) 460-7399 (voice and fax)
> [log in to unmask]