Agile MERODE: a model-driven software engineering method for user-centric and value-based development

Abstract

Agile is often associated with a lack of architectural thinking causing technical debt but has the advantage of user centricity and a strong focus on value. Model-Driven Software Engineering (MDSE) strongly performs for building a quality architecture and code, but lacks focus on user requirements and tends to consider development as a monolithic whole. The combination of Agile and MDSE has been explored, but a convincing integrated method has not been proposed yet. This paper addresses this gap by exploring the specific combination of MERODE -as an example of a proven MDSE method -with Scrum, a reference agile method offering a concrete (sprint-based) life cycle management on the basis of user stories. The method resulting of this integration is called Agile MERODE; it is driven by user stories, themselves associated to Behavior Driven Development scenarios. It allows for domain-driven design and permits fast development from domain models by means of code

Generating multiple conceptual models from behavior-driven development scenarios

“…The model of Robeer et al [47] distinguishes among the role, means, and ends parts of a user story as found in the Connextra template. A unified model of user stories, as per the Connextra template, and scenarios, as per the Gherkin format, is presented by Snoeck and Wautelet [40], based on Tsilionis et al [48] and Wautelet et al [49]. This model describes the concepts of BDD scenarios where a set of scenarios is associated to a user story, which differs from our definition of a BDD scenario as being composed of one user story (described using the Connextra template) and one scenario (described in the Gherkin format).…”

“…The framework could be extended with the concept of events as triggers of actions. The use of the trigger part of BDD scenarios for the creation of models as a basis for model-driven software development has recently been explored by Snoeck and Wautelet [40].…”

“…An interesting avenue is the use of automatically generated models for model-driven software engineering and model-based code generation. In particular, the class diagram and state machine diagram seem to offer opportunities here [40]. We acknowledge that generating models is just a first step in making user stories 'executable', which is a challenge taken up in recent research in Agile software engineering [8,9].…”

Software Requirements Engineering and User Experience Design Modeling of Big Data Analysis using Convolution-Bidirectional Temporal Convolutional Network

“…These days, it has been suggested to use an agent to automate UX testing for certain tasks, like object recognition in game apps [8][9][10].To estimate and predict target variables, machine learning is widely applied in many fields. Reliable datasets are necessary for machine learning models to function well, albeit [11]. For UX based on software needs, there aren't any specific datasets available now.…”