Lessons Learned from the ISBSG Database
by A. Minkiewicz
As corporate subscribers and partners to the International Software Benchmarks Standards Group (ISBSG ), PRICE has access to a wealth of data about software projects.
The ISBSG was formed in 1997 with the mission “To improve the management of IT resources by both business and government through the provision and exploitation of public repositories of software engineering knowledge that are standardized, verified, recent and representative of current technologies.” This database contains detailed information on close to 6000 development and enhancement projects and more than 500 maintenance and support projects. To the best of this author’s knowledge, this database is the largest, most trusted source of publicly available software data that has been vetted and quality checked.
The data covers many industry sectors and types of businesses though it is weak on data in the aerospace and defense industries. Never the less, there are many things we can learn from analysis of this data.
The Development and Enhancement database contains 121 columns of project information for each project submitted. This information includes information identifying the type of business and application, the programming language(s) used, Functional Size of the project in one of many Functional Measures available in the industry (IFPUG, COSMIC, NESMA, etc.), project effort normalized based on the project phases the report contains, Project Delivery Rate (PDR), elapsed project time, etc.
PRICE Systems has recently partnered with ISBSG with licenses to both the data repositories. Although we cannot distribute the data to those without subscriptions, there is no reason we can’t use analysis of this data to provide guidance to users of our software estimation tool, TruePlanning.
One effort focused on developing calibrated software estimation templates for True S based on various scenarios across industry sector, application type, development type (new or enhancement) and language type (3GL,4G). This exercise combined data mining, statistical analysis, and expert judgment.
This paper discusses the methodology used to derive these templates and presents the findings of this research. While the actual analysis is focused on a particular software estimating model, the research, analysis and techniques should inform similar analyses that are tool agnostic.
