A proper characterization of the process of servicing requests for modifications in software systems and identification of the domain-specific aspects of this process are extremely important for software companies since service requests often consume lots of resources. A set of methods and criteria for modeling the servicing process is reviewed in this paper and the results of an extensive empirical investigation based on two commercial applications are presented. Existing software reliability growth models and gamma analysis are adopted to characterize the occurrence and typology of service requests over time. Linear regression is used to describe the learning process. A quantitative assessment is provided on the descriptive and predictive ability of the analyzed models. Since humans are involved in collecting the input data, the sensitivity of the proposed models to human errors is assessed using Monte Carlo simulation.
Succi G, Pedrycz W, Stefanovic M, Russo B (2003). An Investigation on the Occurence of Service Requestes in Commercial Software Applications. EMPIRICAL SOFTWARE ENGINEERING, 8(2), 197-215.
An Investigation on the Occurence of Service Requestes in Commercial Software Applications
Succi G;
2003
Abstract
A proper characterization of the process of servicing requests for modifications in software systems and identification of the domain-specific aspects of this process are extremely important for software companies since service requests often consume lots of resources. A set of methods and criteria for modeling the servicing process is reviewed in this paper and the results of an extensive empirical investigation based on two commercial applications are presented. Existing software reliability growth models and gamma analysis are adopted to characterize the occurrence and typology of service requests over time. Linear regression is used to describe the learning process. A quantitative assessment is provided on the descriptive and predictive ability of the analyzed models. Since humans are involved in collecting the input data, the sensitivity of the proposed models to human errors is assessed using Monte Carlo simulation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.