Towards an Agent-driven Software Architecture Aligned with User Stories

Abstract

Agile principles have taken an increasing importance in the last decades. Software Architecture (SA) definition is perceived as a non-agile practice as it is executed in a top-down manner, reminding waterfall development, and sometimes imposes heavy documentation. This paper proposes to systematically build an agent-oriented SA from a set of User Stories (US), the core artifact to document requirements in agile methodologies. Previous research has allowed to define a unified US meta-model for the generation of templates relating WHO, WHAT and WHY elements. This meta-model's elements define a syntax issued from practitioners templates associated with semantics from Goal Oriented Requirements Engineering frameworks, more precisely i*. With a set of US following the templates of this previous model, the link between the US and SA concepts is systematically studied and a transformation process is proposed. The SA can decline agent behaviors aligned with requirements and organizational behaviors. Moreover, requirements (thus US) are subject to evolution through agile iterations; the SA can evolve with these changes in a semi-automatic manner. We thus argue that the Agent-SA produced with our transformation process contributes to the overall project agility. STRUCTURING USER STORY CONCEPTSThis section exposes the "building blocks" used within this research.

Extracting conceptual models from user stories with Visual Narrator

“…Although [42,65,67] discussed in Sect. 6.1, all introduce methods and tools for extracting models from user stories, these works lack empirical evaluations.…”

“…The authors demonstrate a CASE-tool that automates their approach and allows end users to iteratively improve and extend the output. In another project [67], these same authors propose a method for mapping user stories to agent-based software architecture using i* [17,68]. Similarly, the US2StarTool [42] derives skeletons of i* goal models from user stories.…”

User Stories and Natural Language Processing: A Systematic Literature Review

“…Some studies were authored/co-authored by the same person, indicating the existence of an active research group in this field. [33] Identify ambiguous user stories [34] Define and measure quality factors from user stories [4], [35] Obtain a security defect reporting form from user stories [36] Indicate duplication between user stories [37] Generate model/artifact Generate a test case from user stories [38]- [43] Generate a class diagram from user stories [44], [45] Generate a sequence diagram from user stories [46] Generate a use case diagram from user stories [47]- [49] Generate a use case scenario from user stories [50] Generate a multi-agent system from user stories [51] Generate a source code from user stories [40] Generate a BPMN diagram from user stories [40] Identify the key abstractions To understand the semantic connection in user stories [52]- [54] Identify topics and summarizing user stories [55], [56] Construct a goal model from a set of user stories. [57] Define ontology for user stories [58] Extract the conceptual model of user stories [59], [60] To find the linguistic structure of user stories [61] Prioritizing and estimation of user story complexity [62], [63] Extracting user stories from text [64]- [66] Trace links between model/NL requirements Tracking the development status of user stories from software artifacts [67] Identify the type of dependency of user stories [68] Traceability user stories and software artifact [69]…”

“…The user story can be either input or output of the generated artifact. Fourteen studies reported methods for generating software model/artifacts from user stories, that is, generating a test case from user stories [35][36][37][38][39]64], generating class diagrams from user stories [40,41], generating sequence diagrams from user stories [46], generating a use case diagram from user stories [43][44][45], generating a use case scenario from user stories [50], generating a multi-agent system from user stories [51], generating a source code from user stories [40], and generating BPMN diagrams from user stories [40]. The software artifact generation aims to cut time and cost in software development and avoid inconsistencies, incompleteness, and incorrect requirements and artifact/software models.…”