Towards Model Construction Based on Test Cases and GUI Extraction
The adoption of model-based testing techniques is hindered by the difficulty of creating a test model. Various techniques to automate the modelling process have been proposed, based on software process artefacts or an existing product. This paper outlines
- PDF / 116,338 Bytes
- 6 Pages / 439.37 x 666.142 pts Page_size
- 13 Downloads / 192 Views
bstract. The adoption of model-based testing techniques is hindered by the difficulty of creating a test model. Various techniques to automate the modelling process have been proposed, based on software process artefacts or an existing product. This paper outlines a hybrid approach to model construction, based on two previously proposed methods. The presented approach combines information in pre-existing test cases with a model extracted from the graphical user interface of the product.
Keywords: Model extraction
1
· Model-based testing · Software testing
Introduction
Model-based testing is a testing methodology that automates the generation of tests as well as their execution. In a typical approach, the tester first creates a formal model (such as a state machine) that depicts the behaviour of the system under test (SUT). The model is then explored by an automated tool in order to generate a sequence of actions to be used as a test. Models can also be used to otherwise support the testing process, such as in inspections. A significant drawback of model-based testing is the skill and effort required in modelling. Creating a model that covers all the relevant aspects of the SUT, does so correctly, and is otherwise suitable for test generation, is no small task. Various methods for easing or partially automating the modelling process have been proposed. Models can be generated from different artefacts of the software process, or the artefacts used directly as test models. Suitable candidates include specifications [10] and pre-existing test cases [9,14]. Alternatively, models can be extracted from an existing product, either the source code [4,13,15], the structure and functionality of the graphical user interface (GUI) [1,6,11,12], or other known behaviour [7,8]. Many of these methods also use the results of the generated tests to further hone the model. This paper outlines a hybrid model construction method based on test cases and information extracted from the GUI. Information on the correct behaviour of the SUT in specific states is found in the test cases, and the states can be combined based on information gleaned from the GUI. In this way, weaknesses of c IFIP International Federation for Information Processing 2016 Published by Springer International Publishing AG 2016. All Rights Reserved F. Wotawa et al. (Eds.): ICTSS 2016, LNCS 9976, pp. 225–230, 2016. DOI: 10.1007/978-3-319-47443-4 15
226
A. J¨ aa ¨skel¨ ainen
each method can be compensated for with the strengths of the other. Hopefully, the new method could reduce the effort required to produce a useful test model. The rest of the paper is structured as follows: Sect. 2 presents the two model construction methods that act as the basis for the new approach, which is outlined in Sect. 3. Section 4 considers the potential benefits of the approach. Finally, Sect. 5 concludes the paper with a more general discussion.
2
Automated Model Construction
The approach of this paper builds on two previously presented methods for constructing models to describe the S
Data Loading...
