Next: Results
Up: Metabolic Controls (four parts)
Previous: Specification
After the usual initiatory tests, the calculation of the median was
checked first. For this, three different cases have been
distinguished:
- An even, positive number of arguments was used. In this case, the
result has to be the average of the two elements which are in the
middle of the sorted list of arguments.
- An odd number of arguments was used. In this case, the result has
to be equal to the element in the middle of the sorted list of
arguments.
- An empty list of arguments should not influence the total result
of the statistic.
Some test cases for all three groups have been designed, but no
bug could be found.
The rest of the statistic is only sorting numeric values, and dividing
them into different groups which can be found in the manual
[Holl96] on page 63. Since this are typical tasks for a
database , the majority of the remaining tests done for
this statistic were several volume tests , but
except some discrepancies between the coded product and the
manual which are listed in chapter
on
page
no bugs could be found.
Ingo Melzer
Mon Aug 5 15:12:01 MET DST 1996