Game Development Reference

In-Depth Information

4. Tester's X 130 tests puts the total tests run for the project at 700. Since Z is the current title

holder, X has to have better numbers than Z. Z's new test contribution is 169/700 = 24%.

X's test contribution is 130/700 = 18.5%. X needs to contribute 7% more of the defects

found than Z. Let “x�? be the number of defects X needs to find. Prior to X's defects, the

defect total is 34. When X's defects are found, the new defect total will be 34 + x. X's defect

contribution will be x / (34 + x) and Z's contribution is 9 / (34 + x). Since X's contribution

must be 7% (0.07) higher than Z's, the equation to solve is x / (34 + x) = (9 / (34 + x) ) +

0.07. Eliminate the fractions by multiplying the equation by (34 + x). This gives you x = 9

+ (34 * 0.07) + (x * 0.07). Subtract 0.07x from both sides of the equation and add the con-

stants remaining on the right side to get 0.93x = 11.38. Divide by 0.93 to solve for x, which

is 12.23. Since you can only have whole numbers of defects, X needs to find 13 defects to

grab the “Best Tester�? crown.

5. Some positive aspects of measuring participation and effectiveness: some people will do

better if they know they are being “watched,�? some people will use their own data as

motivation to improve on their numbers during the course of the project, provides a mea-

surable basis for selecting “elite�? testers for promotion or special projects (as opposed to

favoritism for example), testers seeking better numbers may interact more with developers

to find out where to look for defects.

Some negative aspects: effort is required to collect and report this tester data, it can be used

as a “stick�? against certain testers, may unjustly lower the perceived “value�? of testers who

make important contributions in other ways such as mentoring, could lead to jealousy

if one person constantly wins, testers may argue over who gets credit for certain defects

(hinders collaboration and cooperation), some testers will figure a way to exceed at their

individual numbers without really improving the overall test capabilities of the team

(such as choosing easy tests to run).

1. Full combinatorial tables provide all possible combinations of a set of values with each

other. The size of such a table is calculated by multiplying the number of choices being

considered (tested) for each parameter. A pairwise combinatorial table does not have to

incorporate all combinations of every value with all other values. It is “complete�? in the

sense that somewhere in the table there will be at least one instance of any value being

paired up in the same row with any other value. Pairwise tables are typically much smaller

than full combinatorial tables; sometimes hundreds or thousands of times smaller.

2. A parameter represents a function that can be performed by the game or the game player.

Values are the parameter (function) choices that are available, possible or interesting from

a test perspective.

3. Use the template for 7 params to arrive at the table in Figure A.2. The cells with “*�? can have

either a “Yes�? or “No�? value and your table will still be a correct pairwise combinatorial table.

4. In the three new rows, just add one instance of each of the remaining parameters. The

order is not important since they have already been paired with the other parameters in

the original portion of the table. Your new table should resemble the one in Figure A.3.

The cells with “*�? can have either a “Yes�? or “No�? value and your table will still be a correct

pairwise combinatorial table.