The Home Page of the ATeSST group
Department of Information Systems and Computing,
The Analysis, Testing, Slicing, Search and Transformation (ATeSST) group,
which is jointly led by Mark Harman and Rob Hierons,
focuses on formal elements of Software Engineering that contribute
to the development of robust, efficient, reliable, and correct systems.
The work considers the use of formal analysis in:
The work of the ATeSST group is funded, in part, by grants from DaimlerChrysler AG,
the EPSRC and the Leverhulme Trust.
Testing (How can we automatically generate effective and efficient tests?)
Slicing (How can we simplify our code while preserving properties of interest?)
Search (How can we attack Software Engineering problems by representing them
as search problems?)
Transformation (How can we transform our code or specification in order to
aid e.g. testing?)
The ATeSST group is linked to the
GUided Slicing and Targeted Transformation (GUSTT)
GUSTT is an EPSRC funded project concerned with algorithms
and applications of amorphous slicing.
More details are available on the GUSTT page.
Formal Methods and Testing (FORTEST)
FORTEST is an EPSRC funded network that is investigating
the relationships between formal
(and semi-formal) methods and software testing.
FORTEST considers a number of problems such as
the (automatic) generation of efficient and effective
tests from a formal or semi-formal specification
and approaches that combine testing and proof.
More details are available on the
Software Engineering using Metaheuristic Innovative Algorithms (SEMINAL)
This EPSRC funded network aims to apply search techniques,
such as genetic algorithms, simulated annealing
and tabu search to problems typically encountered in the
domain of software engineering.
More details may be found on the
Testability Transformation (TeTra)
TeTra is an EPSRC funded project that
aims to improve software testability using transformations.
The transformations to be used are novel because they need not preserve traditional notion of equivalence. In the conventional sense they are not therefore meaning preserving transformations.
However, they are adequate test set preserving.
The approach is also novel because the transformations are not an end but a means to an end;
once the test data is generated,
the transformed program can be discarded.
A few details are available on the TeTra page.
Evolutionary testing is hard in the presence of side effects, unstructuredness and
flag variables. This project aims to investigate ways in which
program transformation can overcome these barriers to evolutionary program testing.
The Testability Transformation project, funded by Daimler, led to the TeTra project, funded by the EPSRC.
Testing State Based Systems (TESTBASE)
TESTBASE, which is funded by the Leverhulme Trust,
will investigate ways of rewriting state-based specifications in to aid testing.
In particular, it considers the problem of transforming a
state-based specification in order to simplify the problem of
finding a set of feasible paths that, between them, satisfy some
Vada is funded by DaimlerChrysler. Evolutionary testing effectiveness is reduced by large search spaces.
Some variables which are apparently members of the search space, because they are
input variables, do not contribute to the computation under test.
Variable Dependence can identify these variables. This project aims to implement a dependence-based
tool, Vada, for search-space reduction.
Last updated: 30 October 2002.
Disclaimer The contents of this page falls
outside the responsibility of Brunel University.