Index of /~billard/se/cs4320/ex5

      Name                    Last modified       Size  Description


[DIR] Parent Directory        20-Apr-2008 08:49      -  
[TXT] golf                    19-May-2009 09:03    27k  
[TXT] golfscore.txt           19-May-2009 09:03     3k  
[   ] print.bat               19-May-2009 09:03     1k  
[   ] printdoc.bat            19-May-2009 09:03     1k  
[TXT] sample.dat              19-May-2009 09:03     1k  


CS 4320 Software Testing and QA: Ex5 BLACK BOX TEST
===================================================

Platform: UNIX or PC (or paper)

Given: source code

Goal: Use Black Box Test instead of White Box Test.

Black Box Test uses a functional description of the software and then 
applies a method such as:

o equivalence partitioning

o boundary-value analysis

o error guessing

In all cases, Black Box Test treats the software as a box with INPUTS and OUTPUTS,
but does not know, or care, about the inside of the box.

EQUIVALENCE PARTITIONING:

Equivalence partitioning is a systematic process that identifies, on the basis of
whatever information is available, a set of interesting classes of input conditions
to be tested, where each class is representative of (or covers) a large set of other
possible tests. If partitioning is applied to the product under test, the product is
going to behave in much the same way for all members of the class. The aim is to
minimize the number of test cases required to cover these input conditions.

There are two distinct steps. The first is to identify the equivalence classes (ECs)
and the second is to identify a specific test case within the class.

For each external input:

(1) If the input specifies a RANGE of valid values, define one valid EC (within range)
    and two invalid ECs (one below the range and one above the range).
 
    Example: If the input requires a month in the range 1-12, 

        Valid EC: 1..12  Test Case: 4
      Invalid EC: <1     Test Case: -1
      Invalid EC: >12    Test Case: 20

(2) If the input specifies an exact NUMBER (N) of inputs, define one valid EC (with 
    N inputs) and two invalid ECs (<N and >N).

    Example: If the input requires 3 book titles,

       Valid EC: 3       Test Case: 3
     Invalid EC: <3      Test Case: 1
     Invalid EC: >3      Test Case: 5

(3) If the input specifies a SET of valid values, define one valid EC (within set)
    and one invalid EC (outside set).

    Example: If the input requires one of the names TOM, DICK, or HARRY, 

       Valid EC: in  set      Test Case: TOM
     Invalid EC: not in set   Test Case: JOE

(4) If the input specifies a "must be" situation, define one valid EC and one invalid
    EC.

    Example: If the first character of the input must be a digit, 

       Valid EC: 0..9         Test Case: 4
     Invalid EC: not 0..9     Test Case: x

BOUNDARY-VALUE ANALYSIS

Boundary-value analysis (BVA) is a variant and refinement of equivalence partitioning, but
instead of picking an arbitrary member of the class, boundaries are tested because
that usually is a good place to look for defects. For example, suppose N items
are expected as input, a "for loop" in the program might have a "<" sign when it
really needs a "<=" sign. Although this is a black box test (without source code),
BVA can discover this problem.

(1) If the input specifies a RANGE of valid values, 
 
    Example: If the input requires a month in the range 1-12, 

        Valid EC: 1..12  Test Cases: 1,12
      Invalid EC: <1     Test Case: 0
      Invalid EC: >12    Test Case: 13

(2) If the input specifies an exact NUMBER (N) of inputs, 

    Example: If the input requires 3 book titles,

       Valid EC: 3       Test Case: 3
     Invalid EC: <3      Test Case: 2
     Invalid EC: >3      Test Case: 4

Exercises:

Given: a functional design specification

   % more golfscore.txt

a. Run a sample test file (without regard to the test data)

   % golf <sample.dat

b. perform boundary-value analysis

   Do not do Equivalence Partitioning. Just do BVA which is
   just a stricter form.

   Make your test plan first (in BVA.txt). It will look
   something like this:

   (1) Number of golf courses is RANGE 1..4.
                         Test Case    Input File    Output File
     Valid EC: 1..4      1             BVA1.dat       BVA1.out
                         4
   Invalid EC: <1        0
   Invalid EC: >4        5

   Note that BVA1.dat can not possibly test one of the other cases
   listed above. But it might also test 2 golf players. So it will
   be listed again when you make your test plan for the players.
   It is good to reuse a .dat file for VALID ECs but not a good
   idea for INVALID ECs. When bad behavior results, what was the cause
   of it? You should have one .dat per INVALID EC, and it should be
   a simple/clear formulation to demonstrate the bad input.
   This will help isolate the cause and effect.

   Do not start "multiplying" test cases. You don't need to test
   1 golf course VERSUS 2 golfers, VERSUS 10 golfers, etc.

   Do not put any BLANK spaces at the end of your input lines. The input 
   becomes entirely invalid and you can't find out the real behavior.

   Test everything about golfscore.txt even if something other than BVA
   is required. Do not trust that the software works.

   If you do not feel comfortable using a Unix editor (Software Engineers
   should be comfortable in Unix), then SFTP the .dat sources from a PC.
   Note: you may have to call it .txt so it appears nicely in Notepad.

   Given object code, run the test cases:
   % golf <BVA1.dat >BVA1.out

   repeat for BVA2.dat, etc.

c. Look are your *.out files. What errors do you see? The best way to report on
   these errors is to add comments directly into the .out file itself. And include
   the test plan you used to generate this result (from BVA.txt). For example:

   (1) Number of golf courses is RANGE 1..4.
                         Test Case    Input File    Output File
     Valid EC: 1..4      1             BVA1.dat       BVA1.out

   Be sure to give good observations on the results. This will be part of your grade.
   Imagine you are the software engineer or, even better, the ultimate user of the 
   software. Given the specs of the software, what would you like to see happen when 
   given an Invalid EC? How well does the golf program actually do that? That's the 
   purpose of this course.

   After editing your .out files, it's easy to make one concatenation:

   cat *.out >BVA.out

   You can simply turn in BVA.txt (with the entire test plan) and BVA.out (with all
   of your results).

d. Project Notebook (see print.bat)

   README
   golfscore.txt
   sample.dat
   sample.out
   BVA.txt
   BVA1.dat
   BVA2.out
   etc.
   report.txt or include reports in .out files directly

   Email submission: BVA.txt BVA.out (concatenate all)