Databases Reference
In-Depth Information
Fig. 10.3.
Example of input space partitioning structure in the range of
−
15 to 15.
the execution of the software with test data to find errors in the flow of
the control through the software. These test data belong to an input space
which is partitioned into a set of sub domains which belong to a certain
path in the software. The boundary of these domains is obtained by the
predicates in the path condition where a border segment is a section of the
boundary created by a single path condition. Two types of boundary test
points are necessary;
on
and
off
test points. The
on
test points are on the
border within the domain under test, and the
off
test points are outside the
border within an adjacent domain. If the software generates correct results
for all these points then it can be considered that the border locations are
correct.
Domain testing, therefore, is an example of partition testing. In this
testing, first a partition
P
=
{
D
1
∪
D
2
∪···∪
D
n
}
of the input domain
D
is produced. It divides the input space into equivalent domains and it is
assumed that all test data from one domain are expected to be correct
if a selected test data from that domain is shown to be correct. This
form of assumption is called
uniform hypothesis
of partition testing. Each
sub-domain may be affected by two types of faults such as
computation
faults
, which may affect a sub-domain and
domain faults
, which may affect
