Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Model-based functional size measurement
Lavazza L., del Bianco V., Garavaglia C.  Empirical software engineering and measurement (Proceedings of the Second ACM-IEEE International Symposium on Empirical Software Engineering and Measurement, Kaiserslautern, Germany, Oct 9-10, 2008)100-109.2008.Type:Proceedings
Date Reviewed: Dec 9 2008

The function point measure of the functional size of a software set is well known for its use as an input in the estimation of development effort. The authors devised a unified modeling language (UML) model-based approach. This paper introduces a technique to reliably count function points (FPs) from UML models, in accordance with standardized International Function Point Users Group (IFPUG) rules, and without employing a certified counter. The goal of the approach is to improve the process of measuring FPs, and to that end the paper presents mappings between FP elements and UML language constructs. Among the strengths of the paper are clear examples illustrating modeling and counting rules, and an initial empirical validation complemented by an industrial application of the approach.

The approach is model based, as it requires a precise model of requirements obtained using measurement-guided UML modeling. The measurement requirements are essentially homogeneity of models and explicit coverage of all required FP elements. The approach is superior to a posteriori techniques where FP is applied to UML models built without attention to measurement needs. In the central section of the paper, “Measurement-oriented UML Modeling,” using Gym IS case study fragments, the authors explain seven rules about how a UML model of the system to be measured is to be built, and one rule explaining how to count FPs from such models. The following UML diagrams, constrained by the rules, are required: use case, collaboration, class, and sequence. The rules do not extend UML notation, nor do they require information not usually available in detailed UML models. In conclusion, this paper points out that measurement and modeling are not independent, as changes in the model do change its FP count.

The paper is recommended to developers and students already familiar with FPs, and interested in using UML models to calculate FPs.

Reviewer:  Vladan Jovanovic Review #: CR136325 (1002-0168)
Bookmark and Share
 
Product Metrics (D.2.8 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Product Metrics": Date
Communication Metrics for Software Development
Dutoit A., Bruegge B. IEEE Transactions on Software Engineering 24(8): 615-628, 1998. Type: Article
Oct 1 1998
Analyzing Data Sets with Missing Data: An Empirical Evaluation of Imputation Methods and Likelihood-Based Methods
Myrtveit I., Stensrud E., Olsson U. IEEE Transactions on Software Engineering 27(11): 999-1013, 2001. Type: Article
Jul 2 2002
The Optimal Class Size for Object-Oriented Software
El Emam K., Benlarbi S., Goel N., Melo W., Lounis H., Rai S. IEEE Transactions on Software Engineering 28(5): 494-509, 2002. Type: Article
Jan 3 2003
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy