DocumentCode
3373173
Title
An empirical study of the correlations between function point elements [software metrics]
Author
Lokan, Chris J.
Author_Institution
Sch. of Comput. Sci., Australian Defence Force Acad., Canberra, ACT, Australia
fYear
1999
fDate
1999
Firstpage
200
Lastpage
206
Abstract
Researchers and practitioners have noted correlations between the five types of elements in function point analysis. Practitioners exploit them; researchers have raised concerns about them. Two previous studies of the correlations between the elements have agreed in some findings, but differed on others. A large data set is analyzed here to gain further insight into the correlations. Two elements (inputs and logical files) are found to be correlated always, and one element (external files) is found to be generally uncorrelated with the others. Correlations are strongest in new development projects that use 4GLs. These results confirm some findings of previous researchers, and suggest explanations for differences between previous findings
Keywords
software cost estimation; software metrics; 4GLs; data set; empirical study; function point elements; software metrics; Australia; Computer science; Data analysis; Data processing; Size measurement; Software engineering; Software measurement; Software standards;
fLanguage
English
Publisher
ieee
Conference_Titel
Software Metrics Symposium, 1999. Proceedings. Sixth International
Conference_Location
Boca Raton, FL
Print_ISBN
0-7695-0403-5
Type
conf
DOI
10.1109/METRIC.1999.809741
Filename
809741
Link To Document