Problem 1 (4 points): Assume that ACME Software Inc. decides to initiate a measurement effort inside the
software development team. They decide that their software testing (QA) processes are fairly robust and a
great starting point.
What are some measures and metrics that they could use? The business goals are:
Put a simple plan explaining which measures you would choose and how they would help the organization
achieve their business goals.
Answer:
Answer:
Number of revisions per date. To represent this data, we are using the bar chart because we are
representing relative numbers against another relative number that is the date. As the project nears its
completion, the number of revisions per day reduce.
Date Revisions
30 January 2018 420
31 January 2018 360
01 February 2018 270
02 February 2018 330
03 February 2018 70
04 February 2018 5
05 February 2018 350
06 February 2018 90
Revisions
450
400
350
300
250
200
150
100
50
0
Created defects vs Resolved defects. This measure shows the efficiency of the testing team. The more
efficient the testing team is, the higher the reliability and quality of the software product delivered.
For this data we have selected the scattered data plot chart since we are showing the association of the
two data. This can help us in predicting the testing efficiency of the team in future projects.
Created Resolved
Date Defects Defects
30 January
2018 280 244
31 January
2018 280 264
01 February
2018 280 272
02 February
2018 250 244
03 February
2018 250 237
04 February
2018 250 246
05 February
2018 200 192
06 February
2018 200 190
250
200
150
100
50
0
Tuesday,Wednesday,
January 30,
Thursday,
January
2018 31,
February
Friday,
2018February
Saturday,
1, 2018 2,February
Sunday,
2018 February
3,
Monday,
2018 4,February
Tuesday,
2018 5, February
2018 6, 2018
Answer:
Developer User
Effectiveness of inspection and testing. Reliability and usability of the product.
Number of errors per thousand lines of product source code is not the best measurement as the sole
expression of software quality since the number of errors per KLOC would vary for different developers.
Less experienced developers would have more errors per KLOC as compared to more experienced
developers. Also, this measure depends on the effectiveness of inspection and testing team. If there are less
number of errors per KLOC then it might even mean that the testing team is not efficient in finding errors
and not that the software code is correct. This metric might even change because of the complexity of the
software code. Also, this measure is relevant and relies on historical data, meaning that until we have this
measure for another program it is difficult to measure the quality of the present software.
Problem 4 (2 pts): What different product and/or process attributes might the following measure?
1. the number of faults found in program P using a set of test data created specifically for P?
2. the number of faults found in program P using a standard in-house set of test data?
3. the number of faults found in program P by programmer A for one hour.
Answer:
1. Efficiency of test planning team and reliability of the product, stability of the product, cost effectiveness
of the test data for P, complexity of P.
2. Efficiency of the testing process and the testing team, stability of the product.
3. Efficiency, level of experience and the amount of effort required of the programmer A