Information Technology Reference
In-Depth Information
stage. The Preliminary construction screens are not scheduled for many changes be-
yond menu additions. Therefore, the basic screen operation should perform the same
at the end of Final construction. Because the process fl ow redesign will probably
infl uence the way data fi les are built and updated, the original screen operation may
experience new fi le processing overhead. This new fi le processing overhead could
unexpectedly cause the screens to run too slow. If the screens are not regressed for
their performance objectives after the process fl ow redesign has been implemented,
unacceptable performance might be discovered after implementation by DSA.
The second objective is to collect fi rst time performance statistics on the rede-
signed process fl ow when it becomes stable. As with all performance testing, you
should expect the fi rst results to miss the mark. Be prepared from a testing perspec-
tive to characterize as much of the performance problem as many ways as possible
so that the developers can more quickly determine if the performance problem lies
with the redesign code or the redesign itself.
Unlike structural testing in this stage, the Preliminary construction perfor-
mance testing scripts should require very few changes because the overall response
time objectives and strategies for achieving them remain the same in the Final con-
struction stage.
13.6.5 Defect Tracking and Analysis
We continue to plot the defect discoveries for the overall project on the same axis
that we used during Preliminary construction. We are still looking for that peak
around 25 defects per day that we were not able to achieve because, we think, the
basic process fl ow instability interfered. Figure 13.5 shows the results of our defect
tracking as we approach the end of the Final construction stage.
Final
construction
defect peak
35
30
Preliminary
construction
defect peak
25
20
CPI history
DSA project
15
10
5
0
1
6
11
16
21
26
31
36
41
46
51
56
Project days
Figure 13.5
Comparison of CPI defect history with DSA project defect log
Search WWH ::
Custom Search