There are many different approaches to understand and model system requirements. However, systems today tend to be increasingly complex. Agent- and goal-oriented paradigms have been proposed as an alternative to object orientation to cope with these demands. Although it may be intuitive that object-oriented methods could not readily deal with issues such as autonomy, proactiviness and sociality, it is not yet clear to what extent this may be true. Thus, it is necessary to show not only where object orientation might fail, but also where agent/goal orientation still needs improvement. A practical approach that uses a well defined and complex problem producing specifications using agent/goal orientation and object orientation could guide us to understand better the strengths and weaknesses of each approach. That is the main goal of this paper. We use an exemplar proposed in 2001 by Yu and Cysneiros [1] to evaluate both agent/goal orientation and object orientation. For the agent/goal approach we use i*/Tropos, while for object orientation we use UML/RUP. Three teams applied both approaches to the exemplar, producing all the necessary models and answering the evaluation questions provided in the exemplar. We comment in detail on the key findings.
Agent-Oriented; Object-Oriented; Systems Development Methodology; Requirements Engineering; Evaluation.
@inproceedings{wer200501,
author = {Cysneiros, L. M. and Werneck, V. and Amaral, J. and Yu, E.},
title = {Agent/goal Orientation versus Object Orientation for Requirements Engineering: A Practical Evaluation Using an Exemplar},
booktitle = {Proceedings of the WER2005-8th Workshop on Requirements Engineering, Porto - Portugal},
year = {2005},
issn = {2675-0066},
isbn = {972-752-079-0},
doi = {}
}