Information Technology Reference
In-Depth Information
13.5.5 PerformanceTesting
Although there are several data entry screens in the new application, the data entry
screens for use case-03 class roster input have the most critical performance role
with respect to previously identifi ed high business risk. This performance measure-
ment must validate that the new screens do enable the DSA data entry personnel to
complete each certifi cate entry faster than before with the manual system. The target
screen completion speed from our time-motion study at the end of the Analysis stage
is 2.5 min, a fourfold faster completion than the manual system. The fi rst set of test
case PT-03.0 execution results show a certifi cate screen completion time of 4.6 min,
a little over twice as fast as the manual system but not fast enough. The screen devel-
oper and performance tester walk through the data entry screen design and fi nd no
obvious ways to speed up the code. The screen code is really straightforward.
At the tester's suggestion and with DSA's permission, the developer and tester sit
down side-by-side with one of the senior DSA entry staff members and more closely
observe the typing of a dozen or more completion certifi cates. The tester notices
that the DSA employee makes extensive use of the typewriter's Tab key to move the
typewriter to the next fi eld on the certifi cates. Refl ecting on the current operation of
the data entry screen, the developer realizes that the user is forced to move his or her
hands from the keyboard to the mouse in order to position the screen cursor on the
next data entry fi eld. After further discussion, the developer decides to add code to
the data entry screen that provides directed tabbing through the screen in class roster
data fi eld sequence.
When the new tabbing code is added, the test team fi rst performs a regression
test on the data entry screen using test cases FTPOS-03.0 and FTDAT-03.0 to ensure
that the added code does not disturb any of the validated functionality. The test team
then reruns the data entry performance test case PT-03.0 and fi nds that the new, tab-
enable screen can be completed in 3.5 min. This is faster than before but still shy of
the needed fourfold speed increase. The developer and tester returned to the DSA
manual data entry employee and observed some more certifi cate typing.
The tester observes that numeric fi elds on the typewriter do not require the Shift
key, whereas the data entry screen does. Furthermore, there are several screen fi elds
that accepted only digits (class date, instructor code, address zip, and social security
number). The developer agrees to add code to the data entry screen that placed an au-
tomatic NumLock on those number-only screen fi elds to eliminate the user's need to
manually press the NumLock key for those fi elds. Again, the test team regresses the
functional tests and reruns the performance tests. This time the new, tab-enabled,
auto-NumLocked screen is completed in 2.0 min, a fi vefold speed increase!
13.5.6 Defect Tracking and Analysis
As coding is completed and test execution is begun, software defects are discovered
and tracked to correction. The relatively small size of the DCPS allows the
developers to correct defects usually within 36 hours of detection and logging. With
Search WWH ::
Custom Search