An experiment in software sizing with structured analysis metrics

doi:10.1016/0164-1212(91)90053-9

Journal of Systems and Software

Volume 15, Issue 2, May 1991, Pages 159-172

https://doi.org/10.1016/0164-1212(91)90053-9 Get rights and content

Abstract

This paper presents an experiment that was carried out to develop and statistically validate a model that estimates the number of lines of code in a system. The experiment, described using an experimentation frame-work, was conducted at three Canadian organizations.

A linear model is developed for each of the two samples studied. These models are based on measurements taken on data flow diagrams. Statistical analysis reveals that the models produce sound estimates of the actual data (R² = 0.724 and R² = 0.563). Four evaluation criteria show that these models compare favorably to the other software-sizing models found in the literature.

References (33)

V. Côté et al.
Software Metrics: An Overview of Recent Results
J. of Syst. and Soft.
(1988)
B.A. Kitchenham et al.
Software Project Development Cost Estimation
J. of Syst. and Soft.
(1985)
R. Marouane et al.
Economics of Software Project Management in Tunisia, Basic TUCOMO
Inf. and Soft. Tech
(1989)
M. Adler
An Algebra for Data Flow Diagram Process Decomposition
IEEE Trans. on Soft. Eng.
(1988)
J.L. Albin et al.
Collecte et analyse de mesures de logiciel
Techniques et sciences informatiques
(1982)
A.J. Albrecht et al.
Software Function, Source Lines of Code and Development Effort Prediction: A Software Science Validation
IEEE Tr. on Soft. Eng.
(1983)
V.R. Basili et al.
Experimentation in software engineering
IEEE Trans. on Soft. Eng.
(1986)
N.D. Birrell et al.
B.W. Boehm
Software Engineering Economics
(1981)
B. Boehm
Improving Software Productivity
IEEE Computer
(1987)

P. Bourque

Développement d'un modèle statistique d'estimation du nombre de lignes de code fondé sur des métriques de spéification

(1988)

S. Chatterjee et al.

Regression Analysis by Example

(1977)

S.D. Conte et al.

Software Engineering Metrics and Models

(1986)

T. DeMarco

Structured Analysis and System Specification

(1978)

T. DeMarco

Controlling Software Projects

(1982)

DMR

Information System Development Guide—Part 2—Developing the System

(1984)

Cited by (15)

Source code size estimation approaches for object-oriented systems from UML class diagrams: A comparative study
2014, Information and Software Technology
Citation Excerpt :
This will help software managers to make an accurate quotation, efficiently allocate development resources, and effectively schedule development activities [3]. Due to this reason, early code size estimation has been an important research topic in software metrics community for many years [4–16,18,19,22–30,36,37,39,40,54–56,60,61,63,64,69,70]. In the last decade, many early code size estimation approaches for object-oriented systems based on UML (Unified Modeling Language) class diagrams have been proposed [4–16,18,19,25,69,70].
Source code size in terms of SLOC (source lines of code) is the input of many parametric software effort estimation models. However, it is unavailable at the early phase of software development.
We investigate the accuracy of early SLOC estimation approaches for an object-oriented system using the information collected from its UML class diagram available at the early software development phase.
We use different modeling techniques to build the prediction models for investigating the accuracy of six types of metrics to estimate SLOC. The used techniques include linear models, non-linear models, rule/tree-based models, and instance-based models. The investigated metrics are class diagram metrics, predictive object points, object-oriented project size metric, fast&&serious class points, objective class points, and object-oriented function points.
Based on 100 open-source Java systems, we find that the prediction model built using object-oriented project size metric and ordinary least square regression with a logarithmic transformation achieves the highest accuracy (mean MMRE = 0.19 and mean Pred(25) = 0.74).
We should use object-oriented project size metric and ordinary least square regression with a logarithmic transformation to build a simple, accurate, and comprehensible SLOC estimation model.
Software source code sizing using fuzzy logic modeling
2003, Information and Software Technology
Citation Excerpt :
Their experiment was sound and rigorously conducted, although the authors themselves noted that theirs was an assessment of goodness-of-fit modeling rather than of prediction. Bourque and Côté [47] describe an experiment in which they built predictive models of program size using linear regression methods. Employing simple measures extracted from real-world system specifications and designs they were able to generate models that were of comparable accuracy to those produced using far more complex and time-consuming data collection and analysis methods.
Knowing the likely size of a software product before it has been constructed is potentially beneficial in project management: for instance, size can be an important factor in determining an appropriate development/integration schedule, and it can be a significant input in terms of the allocation of personnel and other resources. In this study we consider the applicability of fuzzy logic modeling methods to the task of software source code sizing, using a previously published data set. Our results suggest that, particularly with refinement using data and knowledge, fuzzy predictive models can outperform their traditional regression-based counterparts.
Measuring triggering-interaction complexity on active databases
2001, Information Systems
Distinct software metrics have been proposed for programs. By contrast, metrics for databases have been neglected on the grounds that databases were mere plain files that do not affect considerably information systems maintainability. However, later enhancements on database systems have considerable increase the complexity of the elements kept within the database realm. Such complexity makes metrics a valuable tool to understand, monitor, control, predict and improve software development and maintenance database projects. Triggers are a case in point. Several reports warned about the difficulties to cope with large sets of triggers. Based on the difficulty to ascertain the causes that make a given rule to be triggered, this paper proposes three different metrics for measuring trigger complexity, namely, the triggering potential, the number of anchors and the distance of a trigger. These measures are characterised above the level of the ordinal scale using the measurement theory. Validation of the proposed metrics has been conducted through a set of empirical experiments.
Early size prediction for C and Pascal programs
1996, Journal of Systems and Software
Cost estimation is an important aspect of the management of software development projects. Most cost estimation models require an estimate of the size of the program to be developed. Accurate size estimation is thus critically important. This research relates the size of C and PASCAL programs to information available at successive stages in program development. The data considered are counts of subprograms and variables. Simple models using only these data are shown to be not accurate enough after specification and high-level design. After detailed design, though, they give acceptable estimates for both languages, even across a range of sizes and applications. Estimates based on a weighted variable count (where each variable is weighted by the number of subprograms in which it is referenced) are better than those based on simple variable counts.
Source code size prediction using use case metrics: an empirical comparison with use case points
2017, Innovations in Systems and Software Engineering
Source and test code size prediction a comparison between Use Case Metrics and Objective Class Points
2016, ENASE 2016 - Proceedings of the 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering

View all citing articles on Scopus

¹: Now with the National Bank of Canada, 500 Place d'Armes, Montréal, Canada H2Y 2W3

View full text

An experiment in software sizing with structured analysis metrics

Abstract

J. of Syst. and Soft.

J. of Syst. and Soft.

Inf. and Soft. Tech

An Algebra for Data Flow Diagram Process Decomposition

IEEE Trans. on Soft. Eng.

Collecte et analyse de mesures de logiciel

Techniques et sciences informatiques

Software Function, Source Lines of Code and Development Effort Prediction: A Software Science Validation

IEEE Tr. on Soft. Eng.

Experimentation in software engineering