260 likes | 392 Views
Component characterization: proposal and empirical assessment. Pasquale Ardimento – Phd Candidate Serlab - Dept of Informatics - University of Bari RCost Bari ardimento@di.uniba.it Department of Computation, UMIST Manchester 16th September, 2004. Component-based Software Development.
E N D
Component characterization:proposal and empirical assessment Pasquale Ardimento – Phd Candidate Serlab - Dept of Informatics - University of Bari RCost Bari ardimento@di.uniba.it Department of Computation, UMIST Manchester 16th September, 2004
Component-based Software Development • Component-Based Software Engineering (CBSE) promises to reduce complexity and cost of software processes • but it introduces • difficulty in identifying the most appropriate components, which satisfy specific requirements for a target Components-Based System (CBS) • difficulty in learning the component features
Impact of component choice • The choice of the adequate component is a critical activity because of • asynchronous evolutions between CBS and components integrated in it • various and numerous features of both the organizations and of the systems to develop the most significant aspects of components must be known
Our research • perform empirical studies, aimed at identifying the most significant or relevant characteristics of components • we have identified a set of component characteristics • Adequacy • Adoption cost • Familiarity • Support
Purpose • evaluate the effectiveness of this set wrt a maintainability quality factor • To this end • a quality factor in the CBS has been identified • the effectiveness of component characteristics has been evaluated wrt the maintainability quality factor
Post-Mortem Analysis • Data of three industrial CBSs • SICOD: a CBS supporting the qualification of education and scientific services offered by the University of Bari • DiPNET: a Web-Portal supporting different companies cooperating for the development of a project • Portal: the Web-Portal currently used by University of Bari; it provides a unique point of access to a number of services and information the university offers to its stakeholders • The changes concerned corrective, adaptive and perfective maintenance
Component Characteristics and Metrics... • Adequacy • Functional coverage - the percentage of the CBS functionalities provided by a component wrt to the total number of functionalities the CBS provides • Compliance - the percentage of the CBS functionalities provided by a component wrt to the total number of functionalities the component makes publicly available
...Component Characteristics and Metrics • Adoption cost • Training time - the working time spent for training people involved in the integration of components • Familiarity • the understanding of a component • Support • the level of support provided by the vendor
Quality factor • Mean Maintenance Effort (MME) • MMEj represents the effort spent to satisfy maintenance requests in the j-th component of the CBS where MEi,j is the effort spent for satisfying the i-th maintenance request of the j-th component r is the total number of maintenance requests n is the total number of components in the CBS
Normalization • Values assumed by TT and MME measures are very different among the three projects • these factors had to be normalized • Normalization factor: and so we have:
Investigation Training time Functional Coverage Compliance Familiarity Support Mean Maintenance Effort Research Questions • Is there any correlation between characterization metrics and quality factor metrics? • if so, are the relationships due to specific features of the projects, or are they independent from them?
Mutual Independence of characterization measures Results show that characterization measures are independent from each other
0,047 YES YES 0,008 YES 0,003 Relationship between NMME and characterization measures
Differences among projects … • Are the relationships between characterization metrics and quality factors metrics due to specific features of the projects or are they independent from them? • H0: null hypothesis • some statistically significant difference exists among the three projects with respect to those measures • H1: alternative hypothesis • no statistically significant difference exists among the three projects with respect to those measures
… Differences among projects Correlations are independent from specific features of the three projects
Correlation between component characteristics & MME • Functional Coverage & NMME • The higher is Functional Coverage the higher is the probability that a change in CBS impacts that component • Lesson learned: CBS functionality should be concentrated over a single aspect of the application domain • Training Time & NMME • components vendors propose training period depending on intrinsic difficulty in managing the components • Lesson learned: High training usually indicates the complexity of understanding a component and this implies high maintenance effort • Support & NMME • problems faced during maintenance can be faced making use of the provided support • Lesson learned: a deep knowledge of the component is necessary to avoid the risk that the offered TT is not adequate for comprehending the needed component
Absence of correlation • Compliance • Maintenance effort of a component does not depend on the amount of its usage but on the component itself • Familiarity • During maintenance activity further unfamiliar aspects can be required even when the developer and the maintainer are the same person
Effectiveness of the proposed set of characteristics • The metrics • Functional coverage • Training time • Support resulted to be effective in characterizing the MME quality factor of Components to be maintained in a CBS
Threats • Following threats have been avoided • randomness and irrelevancies of people, tools and processes used within all projects • heterogeneity of experimental subjects • the treatment is unique for all project • Following threats have not been avoided • Reliability of measures • Set of experimental subjects • Set of experimental objects
Future Works… • Since the current study • concerned only three heterogeneous projects • some threats have been poorly considered further investigations are needed to • analyse further characteristics and quality factors, which can be meaningful for choosing components • validate the cause – effect relationship between characteristics and maintenance effort • it will also be necessary to organize metrics collection through more effective techniques and tools than interviews
…Future Works • Following characteristics have still to be investigated • Integration technique • Interface stability • portability • Maintenance effort (a greater detail) • Rating (open source) • Adherence to standards (open source)
Publications on this topic • Bianchi A., Caivano D., Conradi R., Jaccheri L., Torchiano M., Visaggio G., “COTS Products Characterization: Proposal and Empirical Assessment”, in R.Conradi and A.I. Wang (Eds.), Empirical Methods and Studies in Software Engineering – Experiences from ESERNET, Lecture Notes in Computer Science 2765 (2003), 233-255. • P.Ardimento, A.Bianchi, G.Visaggio, “Maintenance-oriented selection of software components”, CSMR2004, Tampere Finland, March: 24-26, 2004 • P.Ardimento, T. Baldassarre, A. Bianchi, “Verso una caratterizzazione delle Componenti Software”, AICA2004, Benevento Italy, September 2004
SICOD • A CBS supporting the qualification of education and scientific services offered by the University of Bari • Components integrated in it are • Oracle 9i • Oracle Internet Application Server 9i • Oracle Report Builder and Crystal Report • Applix iEnterprise • Plumtree Corporate Portal 4.5
DiPNET • a Web-Portal supporting different companies cooperating for the development of a project • Components integrated in it are • SQL Server 7 • SNITZ Forum • DatePicker • MS ADO, COM objects • Internet Information Services (IIS) • MS Index Server (IIXSO) • MS Collaboration Data Objects for NT (CDONTS)
Portal • The Web-Portal currently used by University of Bari • provides a unique point of access to a number of services and information the university offers to its stakeholders • Components integrated in it are • SICOD • SIANAR • Questionnaire • Oracle Portal • Decision Script