Testing the solution

The selection of appropriate test data, including:

• Data that test all the pathways through the algorithm,

• Data that test boundary conditions ‘at’, ‘above’ and ‘below’ values upon which decisions are based,

• Data where the required answer is known and

• Data which is outside the expected values

The need for thorough testing of data

• To find out bugs to debug the program

Testing both algorithms and coded solutions with test data such as:

• Desk checking an algorithm

• Stepping through a coded solution line by line

Desk Checking

Desk checking is the manual/ non-computer way of verifying the programming of algorithms/ logic prior to when they are launched. It helps programmers find bugs and errors which cause the program to run properly. This method is actually used to test pseudo-code before implementing into actual code. However, desk checking can also be used to test & evaluate written code, allowing us to refine or simply the code. 

Desk checking involves taking a sample input to a function you are testing and expressing the outputs of the processes involved. This is done in direct relation to the algorithm code as each variable in it is placed into a table; where the inputs are the first few title columns (these variables can be put into any order but must ensure the output column is after them), then the rest of the variable in alphabetical order (or in order of use) follow, then finally the output variables.

The far left column is always called "Line #", implying that each consecutive row of data corresponds to that line number of the algorithm.

The table is then filled out and is then used to see if the outcome is met with the required output. 

• peer checking

• structured walkthrough