Lines of Code and Man-power estimate for Maxima

I've worked with COCOMO before. The model characteristics can be tweaked
to give "reasonable" estimates, but the model, which you can get cursory
information on from Wikipedia and more detailed data from external links
mentioned at same, require a level of integration between Management,
Software Configuration, and Project Management that you're not likely to
find in the real world. 

The art of SCM is, in most circles, a failure and admitted dead-end, in
no small part because it requires a higher level of knowledge by each of
these groups than one is likely to encounter.

On a personal note, most of the SCM experts I've encountered only know
or understand the "what" and not the "why", making a scaled approach
ludicrous. Most managers of Software Teams I've worked on, even in the
rare occasion that they have programming experience, are so out of touch
with modern Computer Science techniques that, when they talk of SCM,
it's usually because they want it, but again, don't know why or even
what it means. In worst case scenarios, this results in hiring or
placing in charge someone who is an expert SCM that knows "what" but not
"why", resulting in either the extreme case where you have such
burdensome CM policies as to create ridiculous overhead costs as to kill
most projects before they start (and this is touted as a success of SCM)
or to have no meaningful SCM other than to pay someone to take the blame
when things do fail.

Of course, this is my personal opinion based on my limited experiences
(almost exclusively Engineering, and mostly government) with several
different industries, none of which are exclusively software
development, where I suspect you'd find better results.

So I have to ask, why are you doing this estimate? Are you looking to
estimate personal worth? I mean even if you could collect the type of
detailed data you'd need to make something like COCOMO work, you still
couldn't estimate the value of academic contribution created by this
project. You may have much better results (especially since you're in an
academic environment where you can manage the collection of necessary
data) if you apply this data to continuing development. But even in this
case, COCOMO, like many other similar tools is based on statistical
data; attempting to apply it to historical data is always going to be
accurate, but meaningless.

-----Original Message-----
From: maxima-bounces at math.utexas.edu
[mailto:maxima-bounces at math.utexas.edu] On Behalf Of David May
Sent: Thursday, April 22, 2010 9:44 PM
To: Maxima List
Subject: Re: [Maxima] Lines of Code and Man-power estimate for Maxima

Software cost estimation is an black art (possibly a branch of
astrology). For  COCOMO  you specify your model according to various
characteristics of your software project. The output mentions  the
"Basic COCOMO  model" but no details of the model assumptions.
COCOMO is based on an extensive database of projects from
multiple industries including science and medicine. It is not just
"business". Maybe if it was applied correctly, it might even give a
reasonable answer within an order of maginitude.
_______________________________________________
Maxima mailing list
Maxima at math.utexas.edu
http://www.math.utexas.edu/mailman/listinfo/maxima