Customer Reviews

Software Cost Estimation with Cocomo II

5 star:		(3)
4 star:		(4)
3 star:		(0)
2 star:		(0)
1 star:		(1)

 

 

The latest revision of the well known COCOMO constructive cost model is calibrated with a Bayesian approach that balances 161 industry reference points with the determination of software development experts. Software engineers use the COCOMO model to make financial decisions, set project budgets and schedules, negotiate tradeoffs, plan to maintain or upgrade legacy products, and decide where to implement process improvement. The model equations are provided including normative calibration using 5 scale factors and 17post-architecture effort multipliers. The criteria described for assessing your project for scale factors and effort multipliers are fuzzy in some cases (Default to the nominal level when in doubt). The model can accept estimates of either logical lines of code or function points as the primary input parameter. The book is accompanied with a CD that contains the USC implementation of the COCOMO model and some instructional videos. The final 2/3 of the text covers emerging extensions such as object point data, application point data, phase schedule and effort model (COPSEMO), dynamic COCOMO, RAD schedule estimation model (CORADMO), commercial-off-the-shelf integration model (COCOTS), quality estimation (COQUALMO), and productivity estimation (COPROMO). Read "Software Engineering Economics" published in 1981 for additional background information on the COCOMO model.

I was first introduced to the precursor of CoCoMo II when I ftp'd a whitepaper on estimating from the old decwrl site maintained by DEC in the mid 1980s. That led me to Boehm's seminal book, "Software Engineering Economics", ISBN 0138221227. Over time CoCoMo started to show its age, especially in environments with heavy GUI development, so I opted for function points as my sole estimating technique. With experience I discovered that technique alone was useful for estimating scale, but was weak on sizing estimates. Version II of CoCoMo rectifies this, and this book is an exhaustive description of software cost estimating in general and CoCoMo version II in particular.

The contributing authors to this book read like who's who in software estimating, and the contents of this book will not only get you up-to-speed on using CoCoMo II as an estimating tool, but will also explain its inner workings, how to calibrate it, and its inherent strengths and weaknesses. Among the parts of this book that I found especially useful are Chapters 2, Model Definition, and 4, Calibration. These are the keys to understanding and using CoCoMo II to its best advantage. Real world examples are given in Chapter 3, and emerging extensions and future trends provided in chapter 5 and 6 round out the body of this book.

Significant added value is provided in the appendices and on the accompanying CD ROM. The appendices include how to use CoCoMo II in environments ranging from the older waterfall SDLC, RUP, incremental and iterative development, and give project planning considerations and data collection templates. The CD ROM itself contains a full working copy of the 2000 version of USC COCOMO II, complete documentation, and data collection forms, and trial copies of commercial versions of the application.

This book is one that every serious project manager, software engineer and cost estimator should read and keep in their work area.

This book really was written by a committee: there are nine authors. This book is difficult to follow. It does not proceed in a gradual series of steps. Instead, each chapter re-establishes the foundation & context in its own way, often incompletely. This way, each of Dr. Boehm's students had a chance to document his or her thesis topic in the form of a book chapter. Unfortunately, this format does not work for readers who need to start using the data quickly. I'll stick with "Software Engineering Economics", thank you.

COCOMO represents a unique and applicable methodology for process estimates. A goal-driven approach with realistic progress indicators to conducting software estimates represent the way processes must be managed.

The continuity between COCOMOI and COCOMOII is far from being evident. My prime interest is the indentification of the impacts of the Economic Adjustment Factors (EAF) on the economic evolution of a project. An attempt is made to complement COCOMOII with additional tools. There is however an impression of a lack of a common vision.

The book is a major reference in my business development approach. I was privileged to deal with Amazon.com for a fast turn around for the timely delivery of the book at my door.

Bernard Carrier P.Eng.
President OpsPro Inc.

If you not in Dr. Boehm's class at USC, then more than likely your a project manager for a software engineering firm trying to better estimate how much your next project is going to cost.

Either way the book is dry, but the informaiton is great. Its given me insight to CMM and other Software Engineering models.

The COCOMO tool has merit, so does CSCI 577a.

This is great guidence for the people who involved in software cost estimation to deal with the current software engineering issues. It definitely helps to estimate the project's schedule and effort using COCOMO II (Implementation of the model).

We are attempting to use this methodology for planning and review of onboard Aerospace Software, will give an update when we have actually used it in anger!

We are attempting to use this methodology for planning and review of onboard Aerospace Software, will give an update when we have actually used it in anger!