Impact of unified user-story-based modeling on agile methods : aspects on requirements, design and life cycle management

(2017)

Files

SHENG_Thesis_online.pdf
  • Open Access
  • Adobe PDF
  • 9.31 MB
SHENG_Thesis_final.pdf
  • Closed Access
  • Adobe PDF
  • 9.32 MB

Details

Authors
Supervisors
Kolp, Manuel
;
Wautelet, Yves
Abstract
User Stories (US) are the most commonly used requirements artifacts within agile methods such as XP and Scrum. They are written in the form of text of maximum two lines in natural language using prose or following a specific template. Traditionally, they are written down on an index card and posted on a wall or whiteboard for analyzing and monitoring the progress of the project. This performs well when the number of US is limited; but not so well when the number of US is large. In practice, many templates have been proposed with no semantic associated to each syntax used in the templates. This leads to series of issues looking alike different stakeholders interpreting the purpose of the US differently and results in making the whole set difficult to structure and analyze. In the end, this translates in misinterpretation of requirements. This thesis starts with studying existing US templates in order to build a (consistent) unified model for their use with, for each keyword, a syntax and associated semantic. Particularly, this thesis studies US templates following the structure: As [WHO], I want [WHAT], so that [WHY]. Additionally, we also build a derived graphical representation for representing a US set to better analyze dependencies and alternatives of US. This notably allows building a view of the system-to-be on multiple aggregation level. A coarse-grained view can help the development team to better manage the life cycle of the application development by selecting relevant US to be developed within the coming iteration(s). Additionally, a fine-grained view allows dealing with the details of the functions to be developed. This thesis is organized as follows. Part I introduces the work. Part II presents the related literature review—i.e., agile software development, requirements engineering and requirements engineering in agile methods. Part III exposes the unified model for US template which is the pillar of the thesis and is used as a foundation for the other contributions of the thesis. The unified model was constructed based on empirical data collected from the web and scientific literature. The adopted concepts are the Role in the WHO dimension; the Capability, Task, Hard-goal and Soft-goal in the WHAT dimension; and, finally, the Task, Hard-goal and Soft-goal in the WHY dimension. We also provide a well-defined semantic associated to each of these syntaxes; most of them come for the i* framework. Part IV presents the Rationale Tree (RT), a goal-based graphical representation for representing US sets. It indeed allows analyzing the dependencies between US set in a US set—this allows us to better understand the implications in terms of business value offered by their development as well as to get a multiple level view of the system-to-be. Basically, a RT is built out an empirical study on the ability of understanding the constructs of the unified model for tagging a US set and ultimately building a RT out of a US set. The experimentations were conducted with 3 groups: IT Students, Business Students and Researchers. The results have shown that, when tagging a US, identifying the right concept—especially for functional elements where granularity plays an important role—can only be done when evaluating a consistent US set as a whole and thus not on an individual basis. Part V provides an alternative approach for graphically representing a US set based on the (industry adopted) Use-Case model. The latter model is nevertheless intended to only provide a coarse-grained representation of the system-to-be. Finally, Part VI concludes the thesis and discusses future work.
Affiliations

Citations

Heng, S. (2017). Impact of unified user-story-based modeling on agile methods : aspects on requirements, design and life cycle management. https://hdl.handle.net/2078.5/93628