Browse Prior Art Database

A Statistical Approach to Scheduling Software Development

IP.com Disclosure Number: IPCOM000131355D
Original Publication Date: 1978-Dec-01
Included in the Prior Art Database: 2005-Nov-10
Document File: 18 page(s) / 64K

Publishing Venue

Software Patent Institute

Related People

Ware Myers: AUTHOR [+3]

Abstract

[Figure containing following caption omitted: Given certain facts about a project that are known early, this macro-estimating technique generates an expected life- cycle curve of manpower against time.] Is there such a thing as an optimum schedule for a large* software development project? If so, can we find out what it should be early enough in development to be of practical use? Brooks's law' -- adding manpower to a late software project makes it later -- implies that a more relaxed schedule was inevitable from the beginning. Aron,2 reflecting on his experience in observing IBM developments, agrees with Brooks that men and months cannot be freely interchanged. Herd and his co-authors of Doty Associates, Inc.,3 conclude in an Air Force sponsored study that schedule length has a significant impact on project size and programmer productivity. Probably more time than has been scheduled in the past would be required for optimal development. These observations seem to mean that if an appropriate development schedule could be established initially, work of satisfactory quality could be accomplished within the allotted time by the assigned programmers at a high rate of productivity, at a minimum total cost.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 6% of the total text.

Page 1 of 18

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

This record contains textual material that is copyright ©; 1978 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Contact the IEEE Computer Society http://www.computer.org/ (714-821-8380) for copies of the complete work that was the source of this textual material and for all use beyond that as a record from the SPI Database.

A Statistical Approach to Scheduling Software Development

Ware Myers

Computer Staff

(Image Omitted: Given certain facts about a project that are known early, this macro-estimating technique generates an expected life- cycle curve of manpower against time.)

Is there such a thing as an optimum schedule for a large* software development project? If so, can we find out what it should be early enough in development to be of practical use?

Brooks's law' -- adding manpower to a late software project makes it later -- implies that a more relaxed schedule was inevitable from the beginning. Aron,2 reflecting on his experience in observing IBM developments, agrees with Brooks that men and months cannot be freely interchanged. Herd and his co-authors of Doty Associates, Inc.,3 conclude in an Air Force sponsored study that schedule length has a significant impact on project size and programmer productivity. Probably more time than has been scheduled in the past would be required for optimal development.

These observations seem to mean that if an appropriate development schedule could be established initially, work of satisfactory quality could be accomplished within the allotted time by the assigned programmers at a high rate of productivity, at a minimum total cost.

There is reason to believe that in the past many software schedules were established not on the basis of the actual work to be done, but by external factors. In Brooks's much talked about Operating System 360 Project, the initial two- year schedule was determined by the fact that the hardware was planned to be ready in two years. In military systems the date of weapon system availability has sometimes dictated the software development time.3 In commercial systems a product introduction planned for an annual show may crowd soft ware development time.

In the absence of good methods of projecting development time before the work was laid out in detail, software managers have been in a weak post" tion from which to argue against imposed schedules.'The problem has not been ignored, of course -- several new approaches have recently been developed. One, which Lawrence H. Putnam calls macro- estimating,' is the principal subject of this article.

Major approaches

There are two fundamental ways of approaching the task of software estimating. One is to work at the detail level and may be called micro-estimating. The other, macro-estimating, attempts to forecast manpower and schedule from estimators available at the start of design.

Micro-estimating. This conventional techn...