Evaluating a COSMICFFP Measurement Procedure for MultiLayer ObjectOriented

Evaluating a COSMIC-FFP Measurement Procedure for Multi-Layer Object-Oriented Conceptual Schemas Simon Claeys (Master student Ghent University) Geert Poels (Ghent University & Katholieke Universiteit Leuven) Ghent, July 6, 2004 1

Outline n Research context ¨ Development of a COSMIC-FFP measurement procedure for MERODE/OOWS n Problem statement ¨ How n to evaluate this procedure? Possible solution ¨A laboratory experiment to test the procedure according to Abrahão’s general evaluation model for FSM methods Ghent, July 6, 2004 2

Research context – COSMIC-FFP “A more systematic and detailed procedure would provide precise mapping rules for a larger collection of highly specific artifacts, thus diminishing the level of ambiguity when generating the COSMIC-FFP generic software model. Such a procedure would, by definition, be highly dependent on the nature of the artifacts, which, in turn, depends on the software engineering methodology in use in each organization. ” (COSMIC-FFP 2. 2 Measurement Manual, p. 30) Ghent, July 6, 2004 3

Research context – Layered OO Conceptual Model PRESENTATION MODEL Ghent, July 6, 2004 NAVIGATION MODEL OOWS FUNCTIONALITY MODEL MERODE BUSINESS DOMAIN MODEL MERODE 4

Research context – Measurement procedure Mapping of concepts in the MERODE/OOWS meta-models onto the COSMIC-FFP meta-model n COSMIC-FFP 2. 2 mapping rules n ¨ 4 rules for partial business domain model ¨ 4 rules for business domain model ¨ 7 rules for functionality model ¨ 5 rules for navigation model Ghent, July 6, 2004 5

Problem statement: How to evaluate this procedure ? March and Smith’s research framework for IT Research activities Design science Build Evaluate Natural science Theorize Justify Constructs Model Research outputs Method Instantiation "a set of steps (an algorithm or guideline) used to perform a task. Methods are based on a set of underlying constructs (language) and a representation (model) of the solution space. ” (March and Smith 1995, p. 257) Ghent, July 6, 2004 6

Possible solution – Abrahão’s model Inputs Task reduce inputs (increase efficiency) Outputs improve outputs (increase effectiveness) Method n The procedure (rules) should enable the task (applying COSMIC-FFP) ¨ to be performed with less effort ¨ and/or improve the quality of the result ¨ compared to applying COSMIC-FFP without this procedure Ghent, July 6, 2004 7

n Performance => Perceptions => Intentions => Behaviour PERFORMANCE PERCEPTIONS Actual Efficiency Perceived Ease of Use INTENTIONS Intention to Use Actual Effectiveness Perceived Usefulness Actual Efficacy Perceived Efficacy Ghent, July 6, 2004 BEHAVIOUR Actual Usage Adoption in Practice 8

n Operationalize performance properties using ISO/IEC TR 14143 -3: 2003 ¨ ¨ ¨ n Repeatability and reproducability Accuracy Convertibility Discrimination threshold Applicability to functional domains A laboratory experiment is being planned Spring 2005 ¨ Engineers enrolled in a post-graduate Master program in Industrial Management, specialising in ICT (at K. U. Leuven) ¨ FSM with COSMIC-FFP as part of ICT Project Management course ¨ Only a minority has studied MERODE ¨ Ghent, July 6, 2004 9

Possible solution – experimental design n Test principle MERODE/OOWS FUR COSMIC-FFP results Ghent, July 6, 2004 COMPARE Conceptual schema Proposed measurement procedure results 10

n Choices to be made: ¨ Between-subjects versus within-subjects ¨ Choice of application n Reference model (ISO/IEC TR 14143 -4: 2002) Semantically correct MERODE case-study COSMIC-FFP case-studies ¨ Form of the FUR (in the control n Text n ERD, DFD, UI prototype n MERODE/OOWS schema group) ¨ Data analysis and interpretation n How to evaluate accuracy ? (calibration if systematic errors) n How to measure reliability ? n How relevant is measurement time ? n How to interpret perception-based variables ? Ghent, July 6, 2004 11

Ghent, July 6, 2004 12