Boundary Value Testing Chapter 5 Introduction Input domain testing - - PowerPoint PPT Presentation

▶

Feb 01, 2024 589 likes •980 views

Boundary Value Testing Chapter 5 Introduction Input domain testing is the most commonly taught (and perhaps the most commonly used) software testing technique There are a number of approaches to boundary value analysis We will study

SLIDE 1

Boundary Value Testing

Chapter 5

SLIDE 2

BVT–2

Introduction

 Input domain testing is the most commonly taught (and

perhaps the most commonly used) software testing technique

 There are a number of approaches to boundary value

analysis

 We will study some of the limitations of domain testing

SLIDE 3

BVT–3

Program view for boundary analysis

 What is the view we take of a program for boundary

analysis?

SLIDE 4

BVT–4

Program view for boundary analysis – 2

 Many programs can be viewed as a function F that

maps values from a set A (its domain) to values in another set B (its range)

 The input variables of F will have some (possibly

unstated) boundaries:

F(x1,x2) : A "B

a " x1 " b c " x 2 " d

SLIDE 5

BVT–5

What is BVA?

 What is boundary analysis?

SLIDE 6

BVT–6

What is BVA – 2

 For each variable, select five values

 Min

The minimum

 Min+

Slightly above the minimum

 Nom

Nominal

 Max–

Slightly below the maximum

 Max

Maximum

SLIDE 7

BVT–7

Rational for BVA

 What is the rationale for BVA?

SLIDE 8

BVT–8

Rational for BVA – 2

 What is the rationale for BVA?

 Errors occur at boundaries because people

 Mistake logical relations such as mixing < with ≤  Are off by one in counting

 Fence posts and rails

SLIDE 9

BVT–9

Critical assumption

 What is the critical assumption made with boundary

value testing?

SLIDE 10

BVT–10

Critical assumption

 What is the critical assumption made with boundary

value testing?

 Single Fault Assumption

 Failures are rarely the result of the simultaneous

ccurrence of two (or more) faults

SLIDE 11

BVT–11

Single fault assumption

 Based on this assumption

 How are test cases selected when multiple

variables are involved?

SLIDE 12

BVT–12

Single fault assumption – 2

 Generate test cases as such for all i

 Values of all but one variable xi at nominal  xi assumes the four non-nominal values from the slide

Boundary Analysis – 2

 One test case with all nominal values

 What are the number of test cases?

See Figure 5.2

SLIDE 13

BVT–13

Single fault assumption – 3

 Generate test cases as such for all i

 Values of all but one variable xi at nominal  xi assumes the four non-nominal values from the slide

Boundary Analysis – 2

 One test case with all nominal values

 What are the number of test cases?

 #Variables * 4 + 1

See Figure 5.2

SLIDE 14

BVT–14

Two-variable function test cases

<x1nom, x2min> <x1nom, x2min+> <x1nom, x2nom> <x1nom, x2max-> <x1nom, x2max> <x1min, x2nom> <x1min+, x2nom> <x1nom, x2nom> <x1max-, x2nom> <x1max, x2nom>

Apply BVA to the Triangle problem 1 ≤ a ≤ 200 1 ≤ b ≤ 200 1 ≤ c ≤ 200

Redundant tests

SLIDE 15

BVT–15

Advantages

 When does boundary value analysis work well?

SLIDE 16

BVT–16

Advantages – 2

 Independent variables

 Single fault assumption

 Physical quantities  Languages that are not strongly typed

 Why were strongly typed languages developed?

SLIDE 17

BVT–17

Advantages – 3

 Independent variables

 Single fault assumption

 Physical quantities  Languages that are not strongly typed

 Why were strongly typed languages developed?

 To prevent errors easily found with BVA

SLIDE 18

BVT–18

Limitations

 What are the limitations of boundary value analysis?

SLIDE 19

BVT–19

Limitations – 2

 Does not work well for Boolean variables

 Why are these not suitable?

 Does not work well for logical variables

 PIN, transaction type  Why are these not suitable?

 When variables are not independent – i.e. are

dependent

 What example does the textbook give?

 Not that useful for strongly-typed languages

SLIDE 20

BVT–20

Variations of boundary value analysis

 What extensions or variations are made for

boundary value analysis?

 What is the justification for each?

SLIDE 21

BVT–21

Extensions

 Robustness testing  Worst case testing  Robust worst case testing  Special value testing  Random testing

SLIDE 22

Robustness testing

 What is robustness testing?

BVT–22

SLIDE 23

BVT–23

Robustness testing – 2



Add two more values per variable



Max+ Slightly greater than the maximum



Min– Slightly less than the minimum



What is the expected output?



Hopefully error message, system recovers



Implementing these test cases may not be possible



What is the difficulty? See Figure 5.3

SLIDE 24

BVT–24

Robustness testing – 3



Add two more values per variable



Max+ Slightly greater than the maximum



Min– Slightly less than the minimum



What is the expected output?



Hopefully error message, system recovers



Implementing these test cases may not be possible



What is the difficulty?

 Determining the expected output for out of range

values

See Figure 5.3

SLIDE 25

BVT–25

Robustness testing – 4



What are the number of test cases?



When is robust testing mandated?

SLIDE 26

BVT–26

Robustness testing – 5



What are the number of test cases?



#variables * 6 + 1



When is robust testing mandated?



With exception handling

SLIDE 27

Worst case testing

 What is worst case testing?

BVT–27

SLIDE 28

BVT–28

Worst-Case Testing – 2

 Rejects the simple fault assumption and tests all

combinations of values

 Often leads to a large number of test cases with low

bug-finding power

 Why?

 What are the number of test cases?

See Figure 5.4

SLIDE 29

BVT–29

Worst-Case Testing – 3

 Rejects the simple fault assumption and tests all

combinations of values

 Often leads to a large number of test cases with low

bug-finding power

 Why?

 Typically there are few bugs compared to the number

f tests

 What are the number of test cases?

 5 ^ #variables

See Figure 5.4

SLIDE 30

BVT–30

Worst-Case Testing – 4

 What type of testing is better to do in place of worst

case testing?

SLIDE 31

BVT–31

Worst-Case Testing – 5

 What type of testing is better to do in place of worst

case testing?

 Often better to use Special Value Testing

SLIDE 32

Robust worst case testing

 What is robust worst case testing?

BVT–32

SLIDE 33

BVT–33

Robust worst case testing – 2

 Add the values min– and max+ to the possible variable

values

 Now take all combinations of variable values  What are the number of test cases?

See Figure 5.5

SLIDE 34

Special value testing

 What is special value testing?

BVT–34

SLIDE 35

BVT–35

Special value testing – 2

 Use best engineering judgment

 Intuition  Domain knowledge  Experience  Soft spots

SLIDE 36

BVT–36

In class activity



Do exercises 1, 2 and 3

SLIDE 37

BVT–37

Random testing

 Select random values for each variable  How many tests do we make?

SLIDE 38

BVT–38

Random testing – 2

 Select random values for each variable  How many tests do we make?

 Related to the probability of producing every outcome at

least once

 Related to the probability of executing every statement /

path at least once