Domain-Specific Approaches to Software Test Automation (DoSTA)

Domain-Specific Approaches to Software Test Automation (DoSTA)

Workshop at ESEC/FSE 2007, Dubrovnik, Croatia, September 4, 2007

Author Notification Deadline Extended to Monday, June 18

Recent years have witnessed the emergence of domain-specific languages in software development. These languages are tailored to the needs of the domain at hand, thus enabling higher level of abstraction than generic solutions based on standard languages such as UML or Java. These approaches utilize domain knowledge in a way that enables domain experts without programming skills to develop designs that can be converted to optimized implementations by efficient code generators. It is generally agreed that the use of domain-specific solutions can offer vast improvements in productivity.

The aim this workshop is to extend the scope of such domain-specific approaches towards automated software testing at system level. System testing is a context sensitive activity that is by nature more domain specific than program development. However, so far software test automation has failed in its ability to replace manual testing: most of the defects in an average software project are still found manually. Moreover, there are severe test suite maintenance problems associated with test automation systems. Nevertheless, the level of automation in system testing is increasing. Current best practices, such as keyword-driven test frameworks, are considerably more domain-specific than more traditional data-driven ones. In addition, novel model-based solutions, automating not only the test execution but also the test generation, are proving successful in various domains.

However, there are several trade-offs involved concerning for instance the tools, processes, methods, and services. Generic approaches are supported by readily available tools and services while domain-specific solutions need customized design and transformation tools in order to be applicable. Development of generic but customizable tools and processes could be a significant factor in the trade off. Research may provide a resolution to the problem of expensive customization.

This workshop provides a forum for academia, industry, and the research and development community to share their work, exchange experiences and ideas, and discuss challenges and future research directions in the area of domain-specific software test automation. The topics of interest to the workshop include, but are not limited to:

Approach aspects:

• Specific kind of testing approaches for specific domains
• Mixing traditional approaches with new ones
• Domain specific test strategies
• Empirical studies

Methodological aspects:

• Concepts of domain-specific test solutions
• Generation of domain-specific test frameworks

Language aspects:

• Developing testing languages for restricted domains
• Customization of generic languages for domain-specific use (via profiles etc.)
• Transformations between domain-specific languages for test

Tool aspects:

• Experiences in developing domain-specific testing tools
• Customization of generic test or design tools (tool functionality and user experience)
• Design of tools with open interfaces for customization
• Tool presentations

Organizational aspects:

• In-house vs. third party solutions
• Adapting a testing organization for a domain-specific approach
• Customizable testing processes and metrics
• Case studies

Submissions

Paper format: 4 to 6 pages in the ACM conference format available at http://www.acm.org/sigs/pubs/proceed/template.html. A PDF file is uploaded to the workshop submission site by the submission deadline. Electronic proceedings are to be published in conjunction with the proceedings of the main conference by the ACM Digital Library and distributed on a USB-Memory-Stick/CD. Authors of accepted papers must sign an ACM copyright form.