Verifying Web Applications: From Business Level Specifications to Automated Model-Based Testing
This addresses the problem of specification-model disparity for developers in web application testing, but it is incremental as it builds on existing tools like QuickCheck.
The paper tackles the difficulty of adopting model-based testing in industry by proposing an approach to convert Gherkin specifications into models for automated testing, demonstrating its applicability through a case study on the eHealth system in Malta.
One of reasons preventing a wider uptake of model-based testing in the industry is the difficulty which is encountered by developers when trying to think in terms of properties rather than linear specifications. A disparity has traditionally been perceived between the language spoken by customers who specify the system and the language required to construct models of that system. The dynamic nature of the specifications for commercial systems further aggravates this problem in that models would need to be rechecked after every specification change. In this paper, we propose an approach for converting specifications written in the commonly-used quasi-natural language Gherkin into models for use with a model-based testing tool. We have instantiated this approach using QuickCheck and demonstrate its applicability via a case study on the eHealth system, the national health portal for Maltese residents.