DEFRANCO@TOPS20.RADC.AF.MIL (Carl DeFranco) (12/20/89)
I couldn't get this through my mailer back to Dave Prince. Perhaps other
readers would be interested in the reply.
---------------
Date: Wed 20 Dec 89 09:35:32-EST
From: The Mailer Daemon <Mailer@TOPS20.RADC.AF.MIL>
To: DEFRANCO@TOPS20.RADC.AF.MIL
Subject: Message of 20-Dec-89 09:35:22
Message failed for the following:
prince@rutgers.edu: 550 <prince@rutgers.edu>... User unknown
------------
Date: Wed 20 Dec 89 09:35:22-EST
From: Carl DeFranco <DEFRANCO@TOPS20.RADC.AF.MIL>
Subject: Query concerning COCOMO Estimating Model
To: prince@rutgers.edu
Message-ID: <12551614075.14.DEFRANCO@TOPS20.RADC.AF.MIL>
Dave,
I saw your note on a local BB feed. Barry Boehm's big contribution to the
COCOMO model was to fine tune it for Ada. However, I have several sheets
of data about COCOMO from a recent Software Project Management Course I took
and would be happy to fax you copies (or mail as you prefer) of the general
information, including basic equations and cost multiplier factors.
I took the liberty of running a quick calculation through the model, without
assuming any cost factor effects( such as experience, complexity, software
team capability, and a number of machine specific items) i.e. I used 1.0.
FOr a 400K line program, the model indicates about 3,710 man-months of
effort are required, and an allowable schedule of 38 months, with the
corresponding average staffing level of 10+ people. For a more accurate
set of numbers, you will need to plug in the various factors, which can
generate a range of -40% to +150%, depending on impact.
Also, class materials were extracted from a book titled "The Art of
Software Cost Estimation", but I have no author or publication data.
I'm sure someone has put the system into a software module, but I can't
help you much there. Let me know if you'd like the paper data.
Happy Holidays
Carl DeFranco
Rome Air Development Center/COTD
Griffiss AFB, NY 13441-5700
-------
-------
-------