Hardware Reference
In-Depth Information
4.1 Applying Q-PAM
Three faculty members from Tampere University of Technology and four GSD pro-
ject mangers from two industrial companies participated in the assessment workshop
along with the main author. The author of the pattern language introduced a candidate
quality profile in the first evaluation session based on ISO 9126. It was accepted with
some changes after discussion. The main part of the first workshop session was used
for constructing the scenarios and finally 57 scenarios were defined. Those were
prioritized and ten of the most important scenarios were assessed in the second work-
shop session. As an example, the analysis of one of the resulting scenarios is intro-
duced in Table 4.
Table 4.
Example analysis of a scenario
Scenario
S12
An offshore designer decides to decrease the contents of a feature by 50%. In
this way, he/she can get the feature to suit one iteration but the problem is
that he/she doesn't talk with the product manager. This problem should be
visible in two weeks.
Response
A problem need to be solved in GSD as fast as in centralized development.
Quality
Main Factor
Accuracy (Functionality), Time Behaviour (Efficiency)
Pattern
Analysis of Pattern Application
R
N
Iteration Review
The pattern ensures that the change can be found at the latest in the next
Iteration Review.
N
Multi-Level Daily
Meetings
As a result of using this pattern, a project manager might also notice the
change during daily meetings
N
Common
Repositories
and Tools
Common repositories and reports will improve visibility of a project between
different sites and from repositories it is possible to find task lists and reports
e.g about remaining work, in which it is possible to notice the change by this
pattern.
N
Communication
Tools
Communication tools make it easier to clarify change when it has been found.
N
Common
Processes
With Common processes, there can be a risk if there isn't specific process
guidelines to make a decision about making changes and/or all project
members have not been trained well.
R
Result
Some Support:
The implementation of the scenario S12 is supported through four patterns in
the language and one pattern can have a risk.
We illustrate the results of the analysis with a scenario-pattern matrix (Table 5)
where for each scenario the involved patterns are marked with an N (non-risk) or R
(risk).
4.2 Evaluation Indicators
We have computed certain indicator values suggesting problematic scenarios or pat-
terns. These indicators are intended only as hints; the actual conclusions can be made
only after studying the seriousness of each risk separately. We have used the follow-
ing indicators: IR (involvement ratio) = (N+R)/S indicating the potential applicability
scope of the pattern with respect to this set of scenarios, RR (risk ratio) = R/(N+R)
indicating the total degree of risk of the pattern with respect to the scenario set, and SI
(support index) = (N-R)/P indicating the level of support the pattern language pro-
vides for a scenario. Here N and R denote the number of N's and R's in a row/
