CS 217 Software Verification and

Validation

Week 2, Summer 2014

Instructor: Dong Sihttp://www.cs.odu.edu/~dsi

REVIEW OF LAST CLASS

What is (software) testing?

Softeware Testing – definition The process consisting of all life cycle

activities, concerned with planning, preparation and evaluation of software products and related work products to determine:

– that they satisfy specified requirements, – to demonstrate that they are fit for purpose and

– to detect defects

4

Validation & Verification

Validation : Have we built the right software? i.e., do the requirements satisfy the customer? (This is dynamic process for checking and

testing the real product. Software validation always involves with executing the code)

Verification : Have we built the software right? i.e., does it implement the requirements? This is static method for verifying design, code.

Software verification is human based checking of documents and files

Introduction to Software Testing, Edition 2 (Ch 1) © Ammann & Offutt 5

COMPUTER BUG?

What is a computer bug? In 1947 Harvard University was operating a room-sized

computer called the Mark II. – made of vacuum tubes

A moth flew into the computer and was killed by the high voltage. Operators traced an error in the Mark II and taped the bug to log book.

7

Hence, the first computer bug!

I am not making this up :-)

The Term Bug Bug is used informally

Sometimes speakers mean fault, sometimes error, sometimes failure, Incident, problem, Inconsistency … often the speaker doesn’t know what it means !

This class will try to use words that have precise, defined, and unambiguous meanings

Introduction to Software Testing, Edition 2 (Ch 1) © Ammann & Offutt 8

BUG

Introduction to Software Testing, Edition 2 (Ch 1) © Ammann & Offutt 9

Software Fault : A static defect in the software

Software Failure : External, incorrect behavior with respect to the requirements or other description of the expected behavior

Software Error : An incorrect internal state that is the manifestation/expression of some fault

Faults in software are equivalent to design mistakes in hardware.

Software does not degrade.

Software Faults, Errors & Failures

Fault and Failure Example The doctor tries to diagnose the root cause, the

disease– Fault

A patient gives a doctor a list of symptoms– Failures

The doctor may look for anomalous internal conditions (high blood pressure, irregular heartbeat, bacteria in the blood stream)– Errors

Introduction to Software Testing, Edition 2 (Ch 1) © Ammann & Offutt 10

Most medical problems result from external attacks (bacteria, viruses) or physical

degradation as we age.They were there at the beginning and do not

“appear” when a part wears out.

Sources of Problems Requirements Definition: Erroneous, incomplete,

inconsistent requirements.

Design: Fundamental design flaws in the software.

Implementation: Mistakes in chip fabrication, wiring, programming faults, malicious code.

Support Systems: Poor programming languages, faulty compilers and debuggers, misleading development tools.

Sources of Problems (Cont’d)

Inadequate Testing of Software: Incomplete testing, poor verification, mistakes in debugging.

Evolution: Sloppy redevelopment or maintenance, introduction of new flaws in attempts to fix old flaws, incremental escalation to inordinate complexity.

Summary: Why Do We Test Software ?

Introduction to Software Testing, Edition 2 (Ch 1) © Ammann & Offutt 13

A tester’s goal is to eliminate faults as early as possible

• Improve quality• Reduce cost• Preserve customer

satisfaction

Testing main principles

Testing Principles (1) Testing can demonstrate only the presence of

defects and not their absence– Testing can show that defects are present, but cannot

prove that there are no defects. Testing reduces the probability of undiscovered defects remaining in the software but, even if no defects are found, it is not a proof of correctness.

Exhaustive testing is impossible– Exhaustive testing (all combinations of inputs and

preconditions) is not feasible except for trivial cases. Instead of exhaustive testing, risk analysis and priorities should be used to focus testing efforts.

Testing Principles (2)

Early testing is important– Testing activities should start as early as possible in

the software or system development life cycle and should be focused on defined objectives.

Defects are clustering– A small number of modules contain most of the

defects discovered during pre-release testing, or are responsible for the most operational failures.

Testing Principles (3)

Testing is context dependent– Testing is done differently in different contexts. For

example, military software is tested differently from an business site.

Software Testing Process

18

Unit test

Integrationtest

Systemtest

System engineering

Software Design

Code & Implementation

V&V Targets

Software Development Lifecycles

Code and Fix Waterfall Cycle

LOGIC IN COMPUTER SCIENCE

Week 2, topic 1

Motivation Logic became popular in the early 20th century

among philosophers and mathematicians

What constitutes a correct proof in Mathematics?

Some ‘correct’ proofs were later disproved by other mathematicians

Concept of logic helps us to figure out what constitutes a correct argument and what constitutes a wrong argument

21

Motivation Faults (bugs) have been detected in proofs

(programs)

Bugs are hard to detect!

LOGIC enabled mathematicians to point out WHY a proof is wrong, or WHERE in the proof, the reasoning has been faulty.

By symbolizing arguments rather than writing them out in some natural language (which is fraught with ambiguity), checking the correctness of a proof becomes a much more viable task. 22

Motivation

Since the latter half of the 20th century, logic has been used in computer science for various purposes ranging from software validation and verification to theorem-proving.

23

Objective of the class

Understand why logic is important to Software Testing

To prepare the student for using logic as a formal tool in computer science

24

Introduction to Logic Logic is called the CALCULUS of Computer

Science!

LOGIC --- Computer ScienceCALCULUS --- Physical sciences & Engineering

CS areas where we use LOGIC Architecture (logic gates) Software Engineering (Validation & Verification) Programming Languages (Semantics & Logic

Programming) AI (Automatic theorem proving) Algorithms (Complexity) Databases (SQL) 25

History of Logic

Symbolic Logic (500 BC – 19th century)

Algebraic Logic (Mid to late 19th century)

Mathematical Logic (19th century to 20th century)

Logic in Computer Science

26

Fundamental of Logic

Declarative statements

Examples of declarative statements– “A is older than B”– “There is ice in the glass”

– In CIS, describing the data (variables, functions, etc.)

27

Propositions - a statement that is either true or false.

For every proposition p, either p is T or p is F For every proposition p, it is not the case that

p is both T and F

28

Discussion… Which statement is a proposition?

A . Which bird did the cat kill?

B. It is a beautiful day today!

C. William's wife's name is Irene.

D. Give me an A!

29

Discussion… Which statement is a proposition?

A. 2 + 3 = 5

B. 4 / 9

C. 1 + 1 = 3

D. Both A and C

30

Fundamental of Logic We are interested in precise declarative

statements about computer systems and programs

We not only want to specify such statements, but also want to check whether a given program or system fulfills specifications that user needs

31

Fundamental of Logic

5 basic connectives:

– And– Or– Not– If…then…(else)– If and only if

32

Logic in CIS Simple Logic underlies the reasoning in

mathematical statements / programming languages

Objective is to develop codes to model the situations that we encounter

1. Reasoning about situations formally 2. Then constructing arguments/programs

about them that can be executed on a machine

We also want to check whether a computer program or a system satisfies the specifications - TESTING 33

Propositional Logic: Basics Propositional logic describes ways to combine

some true statements to produce other true statements.

If it is proposed that `Jack is taller than John' and

`John can run faster than Jack' are both T =`Jack is taller than John and John can run faster than Jack'.

Propositional logic allows us to formalize such statements.

In concise form: A ^ B34

Propositional Logic: Basics

For CS/CIS, we will restrict our attention to mathematical objects, programs, and data structures in particular.

Statements in a logical language are constructed according to a predefined set of formation rules called ‘syntax rules’.

Same for computer languages

35

C/C++. Java, HTML,…..

All have their own ‘syntax rules’

But the LOGIC behind them are the same !

Propositional Logic: Basics Why English or any other natural language

can’t be used?– Meaning of an English sentence can be ambiguous,

subject to different interpretations depending on the context and implicit assumptions.

– Another important factor is conciseness. Natural languages tend to be wordy, and even fairly simple mathematical statements become exceedingly long (and unclear) when expressed in them.

The logical languages should contain special symbols used for abbreviating syntactical constructs.

Declarative sentences in English string of symbols in CIS

36

Propositional Logic in CIS Compressed but complete encoding of

declarative statements allows us to concentrate on our argumentation

Software is sequence of such declarative statements

Automatic manipulation of such statements, something that machines love to do

37

“Siri” in Apple: “If…Then…” logics in cloud

Propositional Logic

Composition of atomic sentencesp: I won the lottery yesterdayq: I will purchase a lottery ticket todayr: I played a football game yesterday

~ p: Negation. “I did not win the lottery last week”

p v r: Disjunction. The statement is true if at least one of them is true. “I won the lottery or played a football game yesterday.”

38

Propositional Logic

p ^ r: Conjunction. “Yesterday I won the lottery and played a football game.”

p q: Implication. “If I won the lottery last week, then I will purchase a lottery ticket today.” p is called the assumption and q is called conclusion.

39

Propositional Logic p q: Implication, Some interpretations

– p implies q– If p then q– p is a sufficient condition for q– q is a necessary condition for p– q if p– q follows from p– q provided p– q is a consequence of p– q whenever p

40

Natural Deduction Proof

Set of rules which allow us to draw a conclusion by given a set of preconditions

Constructing a proof is much like a programming!

It is not obvious which rules to apply and in what order to obtain the desired conclusion, be careful to choose proof rules!

41

Rules of Natural Deduction Fundamental rule 1 (rule of detachment)

pp q

. . . q

The rule is a valid inference because [p ^ (p q)] q is a tautology!

42

Rules of Natural Deduction

Example: if it is 11:00 o’ clock in Norfolk

if it is 11:00 o’ clock in Norfolk, then it is 11:00 o’ clock in DC

then by rule of detachment, we must conclude:

it is 11:00 o’ clock in DC

43

Rules of Natural Deduction

Fundamental rule 2 (transitive rule)p qq r

. . . p r This is a valid rule of inference because the

implication (p q) ^ (q r) (p r) is a tautology!

44

Rules of Natural Deduction FR 3 (De Morgan’s law)

~(p v q) = (~p) ^ (~q)~(p ^ q) = (~p) v (~q)

FR 4 (Law of contrapositive)p q = (~q ~p)

FR 5 (Double Negation)~(~p) = p

45

Examples of Arguments If a baby is hungry, then the baby cries. If the

baby is not mad, then he does not cry. If a baby is mad, then he has a red face. Therefore, if a baby is hungry, then he has a red face.

Model this problem!!

h: a baby is hungryc: a baby criesm: a baby is madr: a baby has a red face

46

h c~m ~cm r

. . . h r

h cc mm r

. . . h r

Logic is the Skeleton

What remains when arguments are symbolized is the bare logical skeleton

It is this form that enables us to analyze the program / code / software.

Software V&V = Logical proof & Logic error detection

47

CODE COVERAGE

Week 2, topic 2

Definition

Code coverage is a measure used to describe the degree to which the source code of a program is tested by a particular test suite.

A program with high code coverage has been more thoroughly tested and has a lower chance of containing software bugs than a program with low code coverage.

49

Coverage criterias

Function coverage - Has each function (or subroutine) in the program been called?

Statement coverage - Has each statement in the program been executed?

50

√

√

√

Coverage criterias Branch coverage - Has each branch of each

control structure (such as in if and case statements) been executed?

For example, given an if statement, have both the T and F branches been executed?

Another way of saying this is, has every edge in the program been executed?

51

Coverage criterias

Condition coverage - Has each Boolean sub-expression evaluated both to true (T) and false (F) ?

In “A and B”, if sub-expression A is evaluated both to T and F if sub-expression B is evaluated both to T and F

52

Example consider the following C++ function:

If during this execution function 'foo' was called at least once, then function coverage for this function is satisfied.

53

Example consider the following C++ function:

Statement coverage for this function will be satisfied if it was called e.g. as foo(1,1), as in this case, every line in the function is executed including ’z = x;’.

54

Example consider the following C++ function:

Tests calling foo(1,1) and foo(0,1) will satisfy branch coverage because, in the first case, the 2 if conditions are met and z = x; is executed, while in the second case, the first condition (x>0) is not satisfied, which prevents executing z = x;.

55

Example consider the following C++ function:

Condition coverage can be satisfied with tests that call foo(1,1), foo(1,0) and foo(0,0). These are necessary because in the first two cases, (x>0) evaluates to true, while in the third, it evaluates false. At the same time, the first case makes (y>0) true, while the second and third make it false. 56

(x>0) && (y>0) T,F T,F

Condition / branch coverage? Condition coverage does not necessarily imply

branch coverage. For example:

Condition coverage can be satisfied by two tests:

However, this set of tests does not satisfy branch coverage since neither case will meet the if condition.

57