Information Technology Reference
In-Depth Information
could select the equivalence class of ages 1, 50, and 99 and test the same behavior
using 97% fewer test values. The set of values (1, 50, and 99) is said to be an equiva-
lence class of the set of all possible values (1, 2, 3, …, 97, 98, 99).
The example becomes more dramatic if real number values are to be
tested. Consider the same hospital admission system to require a co-pay be-
fore admission (money the patient must pay first before his or her insurance
will pay the rest of the bill). If the co-pay input field allowed from $0.00 (no
co-pay) to $1000.00, then the tester might decide to test all 1,00,001 possible
co-pay values. Alternatively, the tester could select the equivalence class of
co-pay amounts $0.00, $0.99, $1.00, $1.99, $10.00, $10.99, $100.00, $100.99,
$999.00, $999.99, and $1000.00 and test the same behavior using 99.992%
fewer test values.
Please notice that in both examples, test values at the beginning, in the middle,
and end of the data range were chosen as the equivalence class. Correct middle range
value behavior by the software does not necessarily guarantee correct beginning or
end value behavior by the software. Conversely, correct beginning and end value
behavior by the software do not necessarily guarantee correct middle range value
behavior by the software. So, the recommended black box technique for selecting
equivalence class values includes values at the beginning, in the middle, and at
the end of the range. Please notice that the second equivalence class example, co-
pay, also had values for the different possible value lengths (numbers of signifi cant
digits).
7.6.2 Boundary Value Analysis Technique
The boundary value analysis technique extends the analysis of beginning and ending
input value possibilities for an equivalence class. Boundary values are of interest to
testers because a large percentage of functional errors from input and output data
occur on or around these boundaries.
The traditional boundary value analysis begins by identifying the smallest value
increment in a specifi c equivalence class. This smallest value increment is called
the boundary value epsilon . The epsilon is used to calculate
values around
the beginning and ending values in an equivalence class. In the previous example,
the admission age boundary value epsilon is 1 because the input fi eld allows for no
age increment less than a whole year. The admission co-pay boundary value epsi-
lon is $0.01 because the input fi eld allows for no payment increment smaller than a
penny.
In the case of admission age, the beginning boundary test ages would be
/
epsilon around the beginning value 0. The ending
boundary test ages would be 98, 99, and 100, which are
1, 0, and
1, which are
/
epsilon around the
ending value 99. Good boundary value testing technique suggests that you also
test epsilon values around the equivalence class midpoint like 49, 50, and 51 for
the admission age midpoint of 50. The resulting equivalence class with boundary
/
Search WWH ::




Custom Search