Rule of Thumb

This method is used for estimating effort based on practice and previous experience. It is an easily learned and easily applied procedure. This method is not necessarily an accurate or a reliable method as it approximates effort based on past experience without any formal basis. The following rules of thumb, as given in Fairley (1992), can be utilized:

  • (1) Expected productivity in terms of lines of code per programmer month.
  • (2) Quality in terms of errors per thousand lines of code.
  • (3) Percentage distribution of effort and schedule for various activities like design, coding, analysis, etc.

Although many rules of thumb are in use today because they were simple to implement, they can never be replaced by formal estimation methods. Another set of rules of thumb has been developed for forecasting software growth rates during maintenance, optimal enhancement size, and annual software enhancements (Jones, 1996; Smith. 1999).

Estimation by Analogy

In cost estimation methods, a comparison evaluation is made between completed projects and new projects within similar areas of application. From the database for creating the estimate the same sorts of project were chosen in the technique called Estimation of Analogy. Through this analogy approach the cost for the new project was estimated with the assistance of the completed project (Bohem, 1981).

The following are the issues of concern while using the estimation by analogy approach:

  • (1) Right function selection to determine the degree of similarity between target and new project.
  • (2) Relevant projects selection for estimation.
  • (3) Deciding the number of similar projects to be used for estimation.

According to Construx Software Inc., it was found that estimation carried out on the basis of historical data obtained from the organization was more precise when compared to the estimation based on educated guesswork (Construx Software) and rules of thumb. Many organizations do not have sufficient data to use this method of estimation. To overcome this problem, the International Software Benchmarking Standards Group (ISBSG) developed certain metrics for software projects that are found to enhance their benchmarking, productivity, risk analysis, and project estimation most effectively. For performing the evaluation on data repositories several software tools and publications were produced by ISBSG. Based upon metric types certain data will be provided through ISBSG. With the assistance of this data the information regarding past projects in the organization will be gathered and utilized in future estimations. The common entities for which organizations build metrics are

  • (1) Products that are estimated using product metrics.
  • (2) Processes that are estimated using process metrics.
  • (3) Qualities that are predicted using quality metrics.
 
Source
< Prev   CONTENTS   Source   Next >