Software
modelling for automatic test generation, execution, and verification.
The use of model driven test generation is in its infancy and there
are no existing tools that integrate test generation with execution
and verification.
Implementing a test case as a software component (JavaBean). This
is a novel way to improve the maintenance of test suites and facilitate
the incremental development of re-usable testware.
Integration
between test generation and test execution in a distributed and non-deterministic
environment. No currently available tools are able to deal, in
an efficient manner, with test execution in a dynamic, non-deterministic
distributed environment. All available automation technologies assume
the total predictability of the unit under test in order to verify
that a test case has completed successfully.
Feedback
from test execution to further test generation. The use of coverage
analysis and error analysis to enhance the quality of a test suite
is a well-documented technique, but there are no tools to automate
the process. Moreover, the use of coverage feedback is more of a theoretical
rather than an accepted testing practice. AGEDIS will provide a fully
automated feedback path from test execution results to test generation
engine - thus enabling this testing practice to be used in practical
situations with more efficiency.
Use
of exhaustive model traversal and partial search techniques for test
generation. Exhaustive model traversal is only an option if the
number of states is not too large. There are numerous heuristics for
partial search of a finite state machine, and various techniques to
guarantee their effectiveness. AGEDIS will research the most effective
test generation heuristics in the area of distributed software models.
There are likely to be specific techniques for distributed software
that are more appropriate than the generic techniques appearing in
research publications. The consortium will also pioneer new techniques
in this area based on the previous research and technological expertise
of the partners.
Visual
reporting of test results using a test suite browser. Visual presentation
of data is a proven technique for improved productivity. Commercially
available testing tools do not provide visualization of test suites
as a whole in conjunction with their coverage statistics.
An
open architecture that supports standardized and public data exchange
specifications between all major components, to encourage broad acceptance
and exploitation of the tool set by the general testing community
and test tool developers.