Towards knowledge assisted agile requirements evolution

Manish Kumar, Nirav Ajmeri, Smita Ghaisas
2010 Proceedings of the 2nd International Workshop on Recommendation Systems for Software Engineering - RSSE '10  
This paper presents work on a recommendation system for Knowledge assisted Agile Requirements Evolution (K-gileRE). We treat requirements engineering as a special case of knowledge engineering and emphasize the fact that providing a domain knowledge edge can impart agility to the requirements definition exercise. The approach differs from existing agile methods in that it seamlessly incorporates a domain knowledge base into an agile requirements definition framework and explicitly provides to
more » ... quirement analysts, relevant online domain specific recommendations based on underlying ontologies. The framework presents a 'domain knowledge seed' to requirement analysts. The seed provides a view of core features in a given domain and associated knowledge elements such as business processes, rules, policies, partial data models, use cases and test cases,. These in turn are mapped with agile requirements elements such as user stories, features, tasks, product backlog, sprints and prototype plans. The requirement analyst can evolve the seed to suit her specific project needs. As she modifies and evolves the seed specification, she receives domain-specific online recommendations to improve the correctness, consistency and completeness of her requirement specification documents and executable models. Using the domain knowledge seed as a point of departure provides a jump-start to her project. Each exercise of requirements definition thus becomes an evolution from the seed instead of the traditional 'clean slate' Requirements Engineering (RE) that typically starts from the scratch. Hence, the term K-gileRE. We elaborate how K-gileRE helps in practicing the essence of agile doctrines while defining software requirements by providing just-in-time recommendations.
doi:10.1145/1808920.1808924 dblp:conf/icse/KumarAG10 fatcat:aymsaihbrzfvfag3owgwdszsam