UNIT – 1.

INTRODUCTION
1. PURPOSE OF TESTING:

 Testing consumes atleast half of the time and work required to produce a functional
program.
 MYTH: Good programmers write code without bugs. (Its wrong!!!)
 History says that even well written programs still have 1-3 bugs per hundred statements.
 Productivity and Quality in software:
o In production of comsumer goods and other products, every manufacturing stage
is subjected to quality control and testing from component to final stage.
o If flaws are discovered at any stage, the product is either discarded or cycled back
for rework and correction.
o Productivity is measured by the sum of the costs of the material, the rework, and
the discarded componenets, and the cost of quality assurance and testing.
o There is a trade off between quality assurance costs and manufacturing costs: If

UNIT – 1
sufficient time is not spent in quality assurance, the reject rate will be high and so
will be the net cost. If inspection is good and all errors are caught as they occur,
inspection costs will dominate, and again the net cost will suffer.
o Testing and Quality assurance costs for 'manufactured' items can be as low as 2%
in consumer products or as high as 80% in products such as space-ships, nuclear
reactors, and aircrafts, where failures threaten life. Where as the manufacturing
cost of a software is trivial.
o The biggest part of software cost is the cost of bugs: the cost of detecting them, the
cost of correcting them, the cost of designing tests that discover them, and the cost
of running those tests.
o For software, quality and productivity are indistinguishable because the cost of a
software copy is trivial.
 Testing and Test Design are parts of quality assurance should also focus on bug
prevention. A prevented bug is better than a detected and corrected bug.
 Phases in a tester's mental life can be categorised into the following 5 phases:
1. Phase 0: (Until 1956: Debugging Oriented) There is no difference between
testing and debugging. Phase 0 thinking was the norm in early days of software
development till testing emerged as a discipline.
2. Phase 1: (1957-1978: Demonstration Oriented) The purpose of testing here is to
show that software works. Highlighted during the late 1970s. This failed because
the probability of showing that software works 'decreases' as testing
increases. i.e. The more you test, the more likely you'ill find a bug.
3. Phase 2: (1979-1982: Destruction Oriented) The purpose of testing is to show
that software doesnt work. This also failed because the software will never get

. Dept. of CSE Page 2 CREC. Dept. of CSE Page 3

 released as you will find one bug or the other. Also, a bug corrected may also lead Testing Debugging
to another bug.
Testing starts with known conditions, Debugging starts from possibly unknown
4. Phase 3: (1983-1987: Evaluation Oriented) The purpose of testing is not to
uses predefined procedures and has intial conditions and the end can not be
prove anything but to reduce the perceived risk of not working to an acceptable predictable outcomes. predicted except statistically.
value (Statistical Quality Control). Notion is that testing does improve the product
Testing can and should be planned, Procedure and duration of debugging cannot
to the extent that testing catches bugs and to the extent that those bugs are fixed. designed and scheduled. be so constrained.
The product is released when the confidence on that product is high enough. (Note:
This is applied to large software products with millions of code and years of use.) Testing is a demonstration of error or
Debugging is a deductive process.
apparent correctness.
5. Phase 4: (1988-2000: Prevention Oriented) Testability is the factor considered
here. One reason is to reduce the labour of testing. Other reason is to check the Debugging is the programmer's vindication
Testing proves a programmer's failure.
(Justification).
testable and non-testable code. Testable code has fewer bugs than the code that's
hard to test. Identifying the testing techniques to test the code is the main key here. Testing, as executes, should strive to be
Debugging demands intutive leaps,
Test Design: We know that the software code must be designed and tested, but many appear to be predictable, dull, constrained, rigid and
experimentation and freedom.
unaware that tests themselves must be designed and tested. Tests should be properly designed and inhuman.
tested before applying it to the acutal code. Much testing can be done without Debugging is impossible without detailed
Testing is'nt everything: There are approaches other than testing to create better software. design knowledge. design knowledge.
Methods other than testing include:
Testing can often be done by an
6. Inspection Methods: Methods like walkthroughs, deskchecking, formal Debugging must be done by an insider.
outsider.
inspections and code reading appear to be as effective as testing but the bugs
caught donot completely overlap. Much of test execution and design can
Automated debugging is still a dream.
be automated.
7. Design Style: While designing the software itself, adopting stylistic objectives
such as testability, openness and clarity can do much to prevent bugs.
8. Static Analysis Methods: Includes formal analysis of source code during  Function Versus Structure: Tests can be designed from a functional or a structural point
compilation. In earlier days, it is a routine job of the programmer to do that. Now, of view. In functional testing, the program or system is treated as a blackbox. It is
the compilers have taken over that job. subjected to inputs, and its outputs are verified for conformance to specified behaviour.
9. Languages: The source language can help reduce certain kinds of bugs. Functional testing takes the user point of view- bother about functionality and features and
Programmers find new bugs while using new languages. not the program's implementation. Structural testing does look at the implementation
10. Development Methodologies and Development Environment: The details. Things such as programming style, control method, source language, database
development process and the environment in which that methodology is design, and coding details dominate structural testing.
 Both Structural and functional tests are useful, both have limitations, and both target
embedded can prevent many kinds of bugs.
different kinds of bugs. Functional tets can detect all bugs but would take infinite time to
2. DICHOTOMIES: do so. Structural tests are inherently finite but cannot detect all errors even if completely
executed.
 Testing Versus Debugging: Many people consider both as same. Purpose of testing is to  Designer Versus Tester: Test designer is the person who designs the tests where as the
show that a program has bugs. The purpose of testing is to find the error or misconception tester is the one actually tests the code. During functional testing, the designer and tester
that led to the program's failure and to design and implement the program changes that are probably different persons. During unit testing, the tester and the programmer merge
correct the error. into one person.
 Debugging usually follows testing, but they differ as to goals, methods and most  Tests designed and executed by the software designers are by nature biased towards
important psychology. The below tab le shows few important differences between testing structural consideration and therefore suffer the limitations of structural testing.
and debugging.  Modularity Versus Efficiency: A module is a discrete, well-defined, small component of
a system. Smaller the modules, difficult to integrate; larger the modules, difficult to

. Dept. of CSE Page 4 . Dept. of CSE Page 5

 understand. Both tests and systems can be modular. Testing can and should likewise be  ENVIRONMENT:
organised into modular components. Small, independent test cases can be designed to test o A Program's environment is the hardware and software required to make it run. For
independent modules. online systems, the environment may include communication lines, other systems,
 Small Versus Large: Programming in large means constructing programs that consists of terminals and operators.
many components written by many different programmers. Programming in the small is o The environment also includes all programs that interact with and are used to create
what we do for ourselves in the privacy of our own offices. Qualitative and Quantitative the program under test - such as OS, linkage editor, loader, compiler, utility routines.
changes occur with size and so must testing methods and quality criteria. o Because the hardware and firmware are stable, it is not smart to blame the
 Builder Versus Buyer: Most software is written and used by the same organization. environment for bugs.
Unfortunately, this situation is dishonest because it clouds accountability. If there is no  PROGRAM:
separation between builder and buyer, there can be no accountability. o Most programs are too complicated to understand in detail.
 The different roles / users in a system include: o The concept of the program is to be simplified inorder to test it.
1. Builder: Who designs the system and is accountable to the buyer. o If simple model of the program doesnot explain the unexpected behaviour, we may
2. Buyer: Who pays for the system in the hope of profits from providing services. have to modify that model to include more facts and details. And if that fails, we
3. User: Ultimate beneficiary or victim of the system. The user's interests are also may have to modify the program.
guarded by.  BUGS:
4. Tester: Who is dedicated to the builder's destruction. o Bugs are more insidious (deceiving but harmful) than ever we expect them to be.
5. Operator: Who has to live with the builders' mistakes, the buyers' murky (unclear) o An unexpected test result may lead us to change our notion of what a bug is and our
specifications, testers' oversights and the users' complaints. model of bugs.
o Some optimistic notions that many programmers or testers have about bugs are
3. MODEL FOR TESTING: usually unable to test effectively and unable to justify the dirty tests most programs
need.
o OPTIMISTIC NOTIONS ABOUT BUGS:
1. Benign Bug Hypothesis: The belief that bugs are nice, tame and logical.
(Benign: Not Dangerous)
2. Bug Locality Hypothesis: The belief that a bug discovered with in a
component effects only that component's behaviour.
3. Control Bug Dominance: The belief that errors in the control structures (if,
switch etc) of programs dominate the bugs.
4. Code / Data Separation: The belief that bugs respect the separation of code
and data.
5. Lingua Salvator Est: The belief that the language syntax and semantics (e.g.
Structured Coding, Strong typing, etc) eliminates most bugs.
6. Corrections Abide: The mistaken belief that a corrected bug remains
corrected.
7. Silver Bullets: The mistaken belief that X (Language, Design method,
Figure 1.1: A Model for Testing
Above figure is a model of testing process. It includes three models: A model of the representation, environment) grants immunity from bugs.
environment, a model of the program and a model of the expected bugs. 8. Sadism Suffices: The common belief (especially by independent tester) that a
sadistic streak, low cunning, and intuition are sufficient to eliminate most bugs.
Tough bugs need methodology and techniques.

. Dept. of CSE Page 6 . Dept. of CSE Page 7

 9. Angelic Testers: The belief that testers are better at test design than o Sources of Oracles: If every test designer had to analyze and predict the expected
programmers are at code design. behaviour for every test case for every component, then test design would be very
 TESTS: expensive. The hardest part of test design is predicting the expected outcome, but we
o Tests are formal procedures, Inputs must be prepared, Outcomes should predicted, often have oracles that reduce the work. They are:
tests should be documented, commands need to be executed, and results are to be 1. Kiddie Testing: run the test and see what comes out. If you have the outcome
observed. All these errors are subjected to error in front of you, and especially if you have the values of the internal variables,
o We do three distinct kinds of testing on a typical software system. They are: then it is much easier to validate that outcome by analysis and show it to be
1. Unit / Component Testing: A Unit is the smallest testable piece of software correct than it is to predict what the outcome should be and validate your
that can be compiled, assembled, linked, loaded etc. A unit is usually the work prediction.
of one programmer and consists of several hundred or fewer lines of code. Unit 2. Regression Test Suites: Today's software development and testing are
Testing is the testing we do to show that the unit does not satisfy its functional dominated not by the design of new software but by rework and maintenance of
specification or that its implementation structure does not match the intended existing software. In such instances, most of the tests you need will have been
design structure. A Component is an integrated aggregate of one or more run on a previous versions. Most of those tests should have the same outcome
units. Component Testing is the testing we do to show that the component for the new version. Outcome prediction is therefore needed only for changed
does not satisfy its functional specification or that its implementation structure parts of components.
does not match the intended design structure. 3. Purchased Suits and Oracles: Highly standardized software that differ only as
2. Integration Testing: Integration is the process by which components are to implementation often has commercially available test suites and oracles. The
aggregated to create larger components. Integration Testing is testing done to most common examples are compilers for standard languages.
show that even though the componenets were individually satisfactory (after 4. Existing Program: A working, trusted program is an excellent oracle. The
passing component testing), checks the combination of components are typical use is when the program is being rehosted to a new language, operating
incorrect or inconsistent. system, environment, configuration with the intention that the behavior should
3. System Testing: A System is a big component. System Testing is aimed at not change as a result of the rehosting.
revealing bugs that cannot be attributed to components. It includes testing for  IS COMPLETE TESTING POSSIBLE?
performance, security, accountability, configuration sensitivity, startup and o If the objective of the testing were to prove that a program is free of bugs, then
recovery. testing not only would be practically impossible, but also would be theoretically
 Role of Models: The art of testing consists of creating , selecting, exploring, and revising impossible.
models. Our ability to go through this process depends on the number of different models o Three different approaches can be used to demonstrate that a program is
we have at hand and their ability to express a program's behaviour. correct.They are:
 PLAYING POOL AND CONSULTING ORACLES 1. Functional Testing:
o Testing is like playing a pool game. Either you hit the ball to any pocket (kiddie pool)  Every program operates on a finite number of inputs. A complete
or you specify the pocket in advance (real pool). So is the testing. There is kiddie functional test would consists of subjecting the program to all possible
testing and real testing. In kiddie testing, the observed outcome will be considered as input streams.
the expected outcome. In Real testing, the outcome is predicted and documented  For each input the routine either accepts the stream and produces a
before the test is run. correct outcome, accepts the stream and produces an incorrect outcome,
o The tester who cannot make that kind of predictions does not understand the or rejects the stream and tells us that it did so.
program's functional objectives.  For example, a 10 character input string has 280 possible input streams
o Oracles: An oracle is any program, process, or body of data that specifies the and corresponding outcomes, so complete functional testing in this
expected outcome of a set of tests as applied to a tested object. Example of oracle : sense is IMPRACTICAL.
Input/Outcome Oracle - an oracle that specifies the expected outcome for a specified
input.

. Dept. of CSE Page 8 . Dept. of CSE Page 9

  But even theoritically, we can't execute a purely functional test this 2. Correction Cost: What does it cost to correct the bug after it is found? The cost is
way because we don't know the length of the string to which the system the sum of 2 factors: (1) the cost of discovery (2) the cost of correction. These
is responding. costs go up dramatically later in the development cycle when the bug is discovered.
2. Structural Testing: Correction cost also depends on system size.
 The design should have enough tests to ensure that every path through 3. Installation Cost: Installation cost depends on the number of installations: small
the routine is exercised at least once. Right off that's is impossible for a single user program but more for distributed systems. Fixing one bug and
because some loops might never terminate. distributing the fix could exceed the entire system's development cost.
 The number of paths through a small routine can be awesome because 4. Consequences: What are the consequences of the bug? Bug consequences can
each loop multiplies the path count by the number of times through the range from mild to catastrophic.
loop. A reasonable metric for bug importance is
 A small routine can have millions or billions of paths, so total Path Importance= ($) = Frequence * (Correction cost + Installation cost +
Testing is usually IMPRACTICAL. Consequential cost)
3. Formal Proofs of Correctness:  CONSEQUENCES OF BUGS: The consequences of a bug can be measure in terms of

 Formal proofs of correctness rely on a combination of functional and human rather than machine. Some consequences of a bug on a scale of one to ten are:
structural concepts. 1. Mild: The symptoms of the bug offend us aesthetically (gently); a misspelled
 Requirements are stated in a formal language (e.g. Mathematics) and output or a misaligned printout.
each program statement is examined and used in a step of an inductive 2. Moderate: Outputs are misleading or redundant. The bug impacts the system's
proof that the routine will produce the correct outcome for all possible performance.
input sequences. 3. Annoying: The system's behaviour because of the bug is
 The IMPRACTICAL thing here is that such proofs are very expensive dehumanizing. E.g. Names are truncated orarbitarily modified.
and have been applied only to numerical routines or to formal proofs 4. Disturbing: It refuses to handle legitimate (authorized / legal) transactions. The
for crucial software such as system’s security kernel or portions of ATM wont give you money. My credit card is declared invalid.
compilers. 5. Serious: It loses track of its transactions. Not just the transaction itself but the fact
o Each approach leads to the conclusion that complete testing, in the sense of a proof that the transaction occurred. Accountability is lost.
is neither theoretically nor practically possible. 6. Very Serious: The bug causes the system to do the wrong transactions. Instead of
losing your paycheck, the system credits it to another account or converts deposits
 THEORITICAL BARRIERS OF COMPLETE TESTING: to withdrawals.
o "We can never be sure that the specifications are correct" 7. Extreme: The problems aren't limited to a few users or to few transaction types.
o "No verification system can verify every correct program" They are frequent and arbitrary instead of sporadic infrequent) or for unusual
o "We can never be certain that a verification system is correct" cases.
 Not only all known approaches to absoulte demonstrations of correctness impractical, but 8. Intolerable: Long term unrecoverable corruption of the database occurs and the
they are impossible. Therefore, our objective must shift from a absolute proof to a corruption is not easily discovered. Serious consideration is given to shutting the
'suitably convincing' demonstration. system down.
9. Catastrophic: The decision to shut down is taken out of our hands because the
4. CONSEQUENCES OF BUGS: system fails.
10. Infectious: What can be worse than a failed system? One that corrupt other
 IMPORTANCE OF BUGS: The importance of bugs depends on frequency, correction systems even though it doesnot fall in itself ; that erodes the social physical
cost, installation cost, and consequences. environment; that melts nuclear reactors and starts war.
1. Frequency: How often does that kind of bug occur? Pay more attention to the
more frequent bug types.

Dept. of CSE Page 10 . Dept. of CSE Page 11

  FLEXIBLE SEVERITY RATHER THAN ABSOLUTES:  The range is from a few percentage to more than 50%, depending on the
o Quality can be measured as a combination of factors, of which number of bugs and application and environment.
their severity is only one component.  What hurts most about the bugs is that they are the earliest to invade the system
o Many organizations have designed and used satisfactory, quantitative, quality and the last to leave.
metrics. 2. Feature Bugs:
o Because bugs and their symptoms play a significant role in such metrics, as testing  Specification problems usually create corresponding feature problems.
progresses, you see the quality rise to a reasonable value which is deemed to be  A feature can be wrong, missing, or superfluous (serving no useful purpose). A
safe to ship the product. missing feature or case is easier to detect and correct. A wrong feature could
o The factors involved in bug severity are: have deep design implications.
1. Correction Cost: Not so important because catastrophic bugs may be corrected easier  Removing the features might complicate the software, consume more resources,
and small bugs may take major time to debug. and foster more bugs.
2. Context and Application Dependency: Severity depends on the context and the 3. Feature Interaction Bugs:
application in which it is used.  Providing correct, clear, implementable and testable feature specifications is
3. Creating Culture Dependency: Whats important depends on the creators of software not enough.
and their cultural aspirations. Test tool vendors are more sensitive about bugs in their  Features usually come in groups or related features. The features of each group
software then games software vendors. and the interaction of features with in the group are usually well tested.
4. User Culture Dependency: Severity also depends on user culture. Naive users of PC  The problem is unpredictable interactions between feature groups or even
software go crazy over bugs where as pros (experts) may just ignore. between individual features. For example, your telephone is provided with call
5. The software development phase: Severity depends on development phase. Any bugs holding and call forwarding. The interactions between these two features may
gets more severe as it gets closer to field use and more severe the longer it has been have bugs.
around.  Every application has its peculiar set of features and a much bigger set of
unspecified feature interaction potentials and therefore result in feature
5. TAXONOMY OF BUGS: interaction bugs.
Specification and Feature Bug Remedies:
 There is no universally correct way categorize bugs. The taxonomy is not rigid.  Most feature bugs are rooted in human to human communication problems. One
 A given bug can be put into one or another category depending on its history and the solution is to use high-level, formal specification languages or systems.
programmer's state of mind.  Such languages and systems provide short term support but in the long run, does not
 The major categories are: (1) Requirements, Features and Functionality Bugs (2) solve the problem.
Structural Bugs (3) Data Bugs (4) Coding Bugs (5) Interface, Integration and System  Short term Support: Specification languages facilitate formalization of requirements
Bugs (6) Test and Test Design Bugs. and inconsistency and ambiguity analysis.
o REQUIREMENTS, FEATURES AND FUNCTIONALITY BUGS: Various  Long term Support: Assume that we have a great specification language and that can be
categories in Requirements, Features and Functionlity bugs include: used to create unambiguous, complete specifications with unambiguous complete
1. Requirements and Specifications Bugs: testsand consistent test criteria.
 Requirements and specifications developed from them can be incomplete  The specification problem has been shifted to a higher level but not eliminated.
ambiguous, or self-contradictory. They can be misunderstood or impossible to Testing Techniques for functional bugs: Most functional test techniques- that is
understand. those techniques which are based on a behavioral description of software, such as
 The specifications that don't have flaws in them may change while the design is transaction flow testing, syntax testing, domain testing, logic testing and state
in progress. The features are added, modified and deleted. testing are useful in testing functional bugs.
 Requirements, especially, as expressed in specifications are a major source of
expensive bugs.

Dept. of CSE Page 12 . Dept. of CSE Page 13

STRUCTURAL BUGS: Various categories in Structural bugs include: 13. Data-Flow Bugs and Anomalies:
9. Control and Sequence Bugs:  Most initialization bugs are special case of data flow anamolies.
 Control and sequence bugs include paths left out, unreachable code, improper  A data flow anomaly occurs where there is a path along which we expect to do
nesting of loops, loop-back or loop termination criteria incorrect, missing something unreasonable with data, such as using an uninitialized variable,
process steps, duplicated processing, unnecessary processing, rampaging, attempting to use a variable before it exists, modifying and then not storing or
GOTO's, ill-conceived (not properly planned) switches, sphagetti code, and using the result, or initializing twice without an intermediate use.
worst of all, pachinko code.
 One reason for control flow bugs is that this area is amenable (supportive) to DATA BUGS:
theoritical treatment.  Data bugs include all bugs that arise from the specification of data objects, their
 Most of the control flow bugs are easily tested and caught in unit testing. formats, the number of such objects, and their initial values.
 Another reason for control flow bugs is that use of old code especially ALP &  Data Bugs are atleast as common as bugs in code, but they are foten treated as if they
COBOL code are dominated by control flow bugs. didnot exist at all.
 Control and sequence bugs at all levels are caught by testing, especially  Code migrates data: Software is evolving towards programs in which more and more
structural testing, more specifically path testing combined with a bottom line of the control and processing functions are stored in tables.
functional test based on a specification.  Because of this, there is an increasing awareness that bugs in code are only half the
10. Logic Bugs: battle and the data problems should be given equal attention.
 Bugs in logic, especially those related to misundertanding how case statements Dynamic Data Vs Static data:
and logic operators behave singly and combinations  Dynamic data are transitory. Whatever their purpose their lifetime is relatively

 Also includes evaluation of boolean expressions in deeply nested short, typically the processing time of one transaction. A storage object may be
IF-THEN-ELSE constructs. used to hold dynamic data of different types, with different formats, attributes
 If the bugs are parts of logical (i.e. boolean) processing not related to control and residues.
flow, they are characterized as processing bugs.  Dynamic data bugs are due to leftover garbage in a shared resource. This can be

 If the bugs are parts of a logical expression (i.e control-flow statement) which is handled in one of the three ways: (1) Clean up after the use by the user (2)
used to direct the control flow, then they are categorized as control-flow bugs. Common Cleanup by the resource manager (3) No Clean up
11. Processing Bugs:  Static Data are fixed in form and content. They appear in the source code or

 Processing bugs include arithmetic bugs, algebraic, mathematical function database directly or indirectly, for example a number, a string of characters, or a
evaluation, algorithm selection and general processing. bit pattern.
 Examples of Processing bugs include: Incorrect conversion from one data  Compile time processing will solve the bugs caused by static data.

representation to other, ignoring overflow, improper use of grater-than-or-eual  Information, parameter, and control: Static or dynamic data can serve in one of

etc three roles, or in combination of roles: as a parameter, for control, or for information.
 Although these bugs are frequent (12%), they tend to be caught in good unit  Content, Structure and Attributes: Content can be an actual bit pattern, character

testing. string, or number put into a data structure. Content is a pure bit pattern and has no
12. Initialization Bugs: meaning unless it is interpreted by a hardware or software processor. All data bugs
 Initialization bugs are common. Initialization bugs can be improper and result in the corruption or misinterpretation of content. Structure relates to the size,
superfluous. shape and numbers that describe the data object, that is memory location used to store
 Superfluous bugs are generally less harmful but can affect performance. the content. (e.g A two dimensional array). Attributes relates to the specification
 Typical initialization bugs include: Forgetting to initialize the variables before meaning that is the semantics associated with the contents of a data object. (e.g. an
first use, assuming that they are initialized elsewhere, initializing to the wrong integer, an alphanumeric string, a subroutine). The severity and subtlelty of bugs
format, representation or type etc increases as we go from content to attributes because the things get less formal in that
 Explicit declaration of all variables, as in Pascal, can reduce some initialization direction.
problems.
Dept. of CSE Page 14 . Dept. of CSE Page 15
CODING BUGS: 4. Operating System Bugs:
 Coding errors of all kinds can create any of the other kind of bugs.  Program bugs related to the operating system are a combination of
 Syntax errors are generally not important in the scheme of things if the source language hardware architecture and interface bugs mostly caused by a
translator has adequate syntax checking. misunderstanding of what it is the operating system does.
 If a program has many syntax errors, then we should expect many logic and coding  Use operating system interface specialists, and use explicit interface
bugs. modules or macros for all operating system calls.
 The documentation bugs are also considered as coding bugs which may mislead the  This approach may not eliminate the bugs but at least will localize them
maintenance programmers. and make testing easier.
INTERFACE, INTEGRATION, AND SYSTEM BUGS: 5. Software Architecture:
 Various categories of bugs in Interface, Integration, and System Bugs are:  Software architecture bugs are the kind that called - interactive.
1.External Interfaces:  Routines can pass unit and integration testing without revealing such
 The external interfaces are the means used to communicate with the bugs.
world.  Many of them depend on load, and their symptoms emerge only when
 These include devices, actuators, sensors, input terminals, printers, and the system is stressed.
communication lines.  Sample for such bugs: Assumption that there will be no interrupts,
 The primary design criterion for an interface with outside world should Failure to block or un block interrupts, Assumption that memory and
be robustness. registers were initialized or not initialized etc
 All external interfaces, human or machine should employ a protocol.  Careful integration of modules and subjecting the final system to a
The protocol may be wrong or incorrectly implemented. stress test are effective methods for these bugs.
 Other external interface bugs are: invalid timing or sequence 6. Control and Sequence Bugs (Systems Level):
assumptions related to external signals  These bugs include: Ignored timing, Assuming that events occur in a
 Misunderstanding external input or output formats. specified sequence, Working on data before all the data have arrived
 Insufficient tolerance to bad input data. from disc, Waiting for an impossible combination of prerequisites,
2. Internal Interfaces: Missing, wrong, redundant or superfluous process steps.
 Internal interfaces are in principle not different from external interfaces  The remedy for these bugs is highly structured sequence control.
but they are more controlled.  Specialize, internal, sequence control mechanisms are helpful.
 A best example for internal interfaces are communicating routines. 7. Resource Management Problems:
 The external environment is fixed and the system must adapt to it but  Memory is subdivided into dynamically allocated resources such as
the internal environment, which consists of interfaces with other buffer blocks, queue blocks, task control blocks, and overlay buffers.
components, can be negotiated.  External mass storage units such as discs, are subdivided into memory
 Internal interfaces have the same problem as external interfaces. resource pools.
3. Hardware Architecture:  Some resource management and usage bugs: Required resource not
 Bugs related to hardware architecture originate mostly from obtained, Wrong resource used, Resource is already in use, Resource
misunderstanding how the hardware works. dead lock etc
 Examples of hardware architecture bugs: address generation error, i/o  Resource Management Remedies: A design remedy that prevents
device operation / instruction error, waiting too long for a response, bugs is always preferable to a test method that discovers them.
incorrect interrupt handling etc.  The design remedy in resource management is to keep the resource
 The remedy for hardware architecture and interface problems is two structure simple: the fewest different kinds of resources, the fewest
fold: (1) Good Programming and Testing (2) Centralization of hardware pools, and no private resource management.
interface software in programs written by hardware interface
specialists.

. Dept. of CSE Page 16 . Dept. of CSE Page 17

 8. Integration Bugs: Assemblers, loaders, compilers are developed to reduce the incidence of
 Integration bugs are bugs having to do with the integration of, and with programming and operation errors. Test execution bugs are virtually eliminated
the interfaces between, working and tested components. by various test execution automation tools.
 These bugs results from inconsistencies or incompatibilities between 4. Test Design Automation: Just as much of software development has been
components. automated, much test design can be and has been automated. For a given
 The communication methods include data structures, call sequences, productivity rate, automation reduces the bug count - be it for software or be it for
registers, semaphores, communication links and protocols results in tests.
integration bugs.
 The integration bugs do not constitute a big bug category(9%) they are
expensive category because they are usually caught late in the game and
because they force changes in several components and/or data
structures.
9.System Bugs:
 System bugs covering all kinds of bugs that cannot be ascribed to a
component or to their simple interactions, but result from the totality of
interactions between many components such as programs, data,
hardware, and the operating systems.
 There can be no meaningful system testing until there has been
thorough component and integration testing.
 System bugs are infrequent(1.7%) but very important because they are
often found only after the system has been fielded.

TEST AND TEST DESIGN BUGS:

 Testing: testers have no immunity to bugs. Tests require complicated scenarios and
databases.
 They require code or the equivalent to execute and consequently they can have
bugs.
 Test criteria: if the specification is correct, it is correctly interpreted and
implemented, and a proper test has been designed; but the criterion by which the
software's behavior is judged may be incorrect or impossible. So, a proper test
criteria has to be designed. The more complicated the criteria, the likelier they are
to have bugs.
Remedies: The remedies of test bugs are:
1. Test Debugging: The first remedy for test bugs is testing and debugging the tests.
Test debugging, when compared to program debugging, is easier because tests,
when properly designed are simpler than programs and donot have to make
concessions to efficiency.
2. Test Quality Assurance: Programmers have the right to ask how quality in
independent testing is monitored.
3. Test Execution Automation: The history of software bug removal and
prevention is indistinguishable from the history of programming automation aids.

. Dept. of CSE Page 18 . Dept. of CSE Page 19

 FLOWGRAPHS AND PATH TESTING  A process has one entry and one exit. It can consists of a single statement
or instruction, a sequence of statements or instructions, a single
entry/exit subroutine, a macro or function call, or a sequence of these.
1. BASICS OF PATH TESTING: 2. Decisions:
 A decision is a program point at which the control flow can diverge.
 PATH TESTING:  Machine language conditional branch and conditional skip instructions
o Path Testing is the name given to a family of test techniques based on judiciously are examples of decisions.
selecting a set of test paths through the program.  Most of the decisions are two-way but some are three way branches in
o If the set of paths are properly chosen then we have achieved some measure of test control flow.
thoroughness. For example, pick enough paths to assure that every source 3. Case Statements:
statement has been executed at least once.  A case statement is a multi-way branch or decisions.
o Path testing techniques are the oldest of all structural test techniques.  Examples of case statement are a jump table in assembly language, and
o Path testing is most applicable to new software for unit testing. It is a structural the PASCAL case statement.
technique.  From the point of view of test design, there are no differences between
o It requires complete knowledge of the program's structure. Decisions and Case Statements
o It is most often used by programmers to unit test their own code. 4. Junctions:
o The effectiveness of path testing rapidly deteriorates as the size of the software  A junction is a point in the program where the control flow can merge.
aggregate under test increases.  Examples of junctions are: the target of a jump or skip instruction in ALP,
 THE BUG ASSUMPTION: a label that is a target of GOTO.
o The bug assumption for the path testing strategies is that something has gone
wrong with the software that makes it take a different path than intended.
o As an example "GOTO X" where "GOTO Y" had been intended.
o Structured programming languages prevent many of the bugs targeted by path
testing: as a consequence the effectiveness for path testing for these languages is
reduced and for old code in COBOL, ALP, FORTRAN and Basic, the path testing
is indespensable.
 CONTROL FLOW GRAPHS:
o The control flow graph is a graphical representation of a program's control
structure. It uses the elements named process blocks, decisions, and junctions.
o The flow graph is similar to the earlier flowchart, with which it is not to be
confused.
o Flow Graph Elements:A flow graph contains four different types of elements. (1)
Process Block (2) Decisions (3) Junctions (4) Case Statements
1. Process Block:
 A process block is a sequence of program statements uninterrupted by
either decisions or junctions.
 It is a sequence of statements such that if any one of statement of the
block is executed, then all statement thereof are executed.
 Formally, a process block is a piece of straight line code of one statement
or hundreds of statements.
Figure 1.1: Flowgraph Elements

. Dept. of CSE Page 20 . Dept. of CSE Page 21

CONTROL FLOW GRAPHS Vs FLOWCHARTS:
 A program's flow chart resembles a control flow graph.
 In flow graphs, we don't show the details of what is in a process block.
 In flow charts every part of the process block is drawn.
 The flowchart focuses on process steps, where as the flow graph focuses on control flow
of the program.
 The act of drawing a control flow graph is a useful tool that can help us clarify the control
flow and data flow issues.
NOTATIONAL EVOULTION:
The control flow graph is simplified representation of the program's structure.
The notation changes made in creation of control flow graphs:
 The process boxes weren't really needed. There is an implied process on
every line joining junctions and decisions.
 We don't need to know the specifics of the decisions, just the fact that there is
a branch.
 The specific target label names aren't important-just the fact that they exist.
So we can replace them by simple numbers.
To understand this, we will go through an example (Figure 1.2) written in a FORTRAN
like programming language called Programming Design Language (PDL). The program's
corresponding flowchart (Figure 1.3) and flowgraph (Figure 1.4) were also provided below for
better understanding. Figure 1.3: One-to-one flowchart for example
program in Figure1.2
The first step in translating the program to a flowchart is shown in Figure 1.3, where we have
the typical one-for-one classical flowchart. Note that complexity has increased, clarity has
decreased, and that we had to add auxiliary labels (LOOP, XX, and YY), which have no actual
program counterpart. In Figure 1.4 we merged the process steps and replaced them with the single
process box. We now have a control flowgraph. But this representation is still too busy. We
simplify the notation further to achieve Figure 1.5, where for the first time we can really see what
the control flow looks like.

Figure 1.2: Program Example (PDL) Figure 1.4: Control Flowgraph for example in Figure 1.2
Dept. of CSE Page 22 Dept. of CSE Page 23
 Linked List representation of Flow Graph:

Figure 1.5: Simplified Flowgraph Notation

Figure 1.6: Even Simplified Flowgraph Notation

The final transformation is shown in Figure 1.6, where we've dropped the node numbers to
achieve an even simpler representation. The way to work with control flowgraphs is to use the Figure 1.7: Linked List Control Flowgraph Notation
simplest possible representation - that is, no more information than you need to correlate back to
the source program or PDL. FLOWGRAPH - PROGRAM CORRESPONDENCE:

LINKED LIST REPRESENTATION: A flow graph is a pictorial representation of a program and not the program itself, just as a
topographic map.
Although graphical representations of flowgraphs are revealing, the details of the control flow You cant always associate the parts of a program in a unique way with flowgraph parts because
inside a program they are often inconvenient. many program structures, such as if-then-else constructs, consists of a combination of decisions,
In linked list representation, each node has a name and there is an entry on the list for each link in junctions, and processes.
the flow graph. only the information pertinent to the control flow is shown. The translation from a flowgraph element to a statement and vice versa is not always unique. (See
Figure 1.8)

Dept. of CSE Page 24 Dept. of CSE Page 25

 FUNDAMENTAL PATH SELECTION CRITERIA:
There are many paths between the entry and exit of a typical routine.
Every decision doubles the number of potential paths. And every loop multiplies the number of
potential paths by the number of different iteration values possible for the loop.
Defining complete testing:
1. Exercise every path from entry to exit
2. Exercise every statement or instruction at least once
3. Exercise every branch and case statement, in each direction at least once
If prescription 1 is followed then 2 and 3 are automatically followed. But it is impractical for most
routines. It can be done for the routines that have no loops, in which it is equivalent to 2 and 3
prescriptions.
EXAMPLE:Here is the correct version.

Figure 1.8: Alternative Flowgraphs for same logic (Statement

"IF (A=0) AND (B=1) THEN . . .").

An improper translation from flowgraph to code during coding can lead to bugs, and improper For X negative, the output is X + A, while for X greater than or equal to zero, the output is X + 2A.
translation during the test design lead to missing test cases and causes undiscovered bugs. Following prescription 2 and executing every statement, but not every branch, would not reveal
FLOWGRAPH AND FLOWCHART GENERATION: the bug in the following incorrect version:
Flowcharts can be
1. Handwritten by the programmer.
2. Automatically produced by a flowcharting program based on a mechanical analysis of the
source code.
3. Semi automatically produced by a flow charting program based in part on structural analysis
of the source code and in part on directions given by the programmer.
There are relatively few control flow graph generators. A negative value produces the correct answer. Every statement can be executed, but if the test
cases do not force each branch to be taken, the bug can remain hidden. The next example uses a
PATH TESTING - PATHS, NODES AND LINKS: test based on executing each branch but does not force the execution of all statements:
Path: a path through a program is a sequence of instructions or statements that starts at an entry,
junction, or decision and ends at another, or possibly the same junction, decision, or exit. A path
may go through several junctions, processes, or decisions, one or more times.
 Paths consists of segments.
 The segment is a link - a single process that lies between two nodes.
 A path segment is succession of consecutive links that belongs to some path.
 The length of path measured by the number of links in it and not by the number of the
instructions or statements executed along that path.
 The name of a path is the name of the nodes along the path.
The hidden loop around label 100 is not revealed by tests based on prescription 3 alone because
no test forces the execution of statement 100 and the following GOTO statement. Furthermore,

Dept. of CSE Page 26 Dept. of CSE Page 27

label 100 is not flagged by the compiler as an unreferenced label and the subsequent GOTO does 
For structured software, branch testing and therefore branch coverage
not refer to an undefined label. strictly includes statement coverage.
 We denote branch coverage by C2.
A Static Analysis (that is, an analysis based on examining the source code or structure) cannot
Commonsense and Strategies:
determine whether a piece of code is or is not reachable. There could be subroutine calls with
 Branch and statement coverage are accepted today as the minimum
parameters that are subroutine labels, or in the above example there could be a GOTO that
targeted label 100 but could never achieve a value that would send the program to that label. mandatory testing requirement.
 The question "why not use a judicious sampling of paths?, what is wrong
Only a Dynamic Analysis (that is, an analysis based on the code's behavior while running - with leaving some code, untested?" is ineffectual in the view of common
which is to say, to all intents and purposes, testing) can determine whether code is reachable or sense and experience since: (1.) Not testing a piece of a code leaves a
not and therefore distinguish between the ideal structure we think we have and the actual, buggy residue of bugs in the program in proportion to the size of the untested code
structure.
and the probability of bugs. (2.) The high probability paths are always
PATH TESTING CRITERIA: thoroughly tested if only to demonstrate that the system works properly.
Any testing strategy based on paths must at least both exercise every instruction and take  Which paths to be tested? You must pick enough paths to achieve C1+C2.
branches in all directions. The question of what is the fewest number of such paths is interesting to the
A set of tests that does this is not complete in an absolute sense, but it is complete in the sense that designer of test tools that help automate the path testing, but it is not crucial
anything less must leave something untested. to the pragmatic (practical) design of tests. It is better to make many simple
So we have explored three different testing criteria or strategies out of a potentially infinite family paths than a few complicated paths.
of strategies. LOOPS:
1. Path Testing (Pinf):  Cases for a single loop: A Single loop can be covered with two cases: Looping and

 Execute all possible control flow paths through the program: typically, this Not looping. But, experience shows that many loop-related bugs are not discovered
is restricted to all possible entry/exit paths through the program. by C1+C2. Bugs hide themselves in corners and congregate at boundaries - in the
 If we achieve this prescription, we are said to have achieved 100% path cases of loops, at or around the minimum or maximum number of times the loop can
coverage. This is the strongest criterion in the path testing strategy family: be iterated. The minimum number of iterations is often zero, but it need not be.
it is generally impossible to achieve. CASE 1: Single loop, Zero minimum, N maximum, No excluded values
1. Statement Testing (P1): 1. Try bypassing the loop (zero iterations). If you can't, you either have a
 Execute all statements in the program at least once under some test. If we bug, or zero is not the minimum and you have the wrong case.
do enough tests to achieve this, we are said to have achieved 100% 2. Could the loop-control variable be negative? Could it appear to specify a
statement coverage. negative number of iterations? What happens to such a value?
 An alternate equivalent characterization is to say that we have achieved 100% 3. One pass through the loop.
node coverage. We denote this by C1. 4. Two passes through the loop.
 This is the weakest criterion in the family: testing less than this for new 5. A typical number of iterations, unless covered by a previous test.
software is unconscionable (unprincipled or can not be accepted) and 6. One less than the maximum number of iterations.
should be criminalized. 7. The maximum number of iterations.
2. Branch Testing (P2): 8. Attempt one more than the maximum number of iterations. What
 Execute enough tests to assure that every branch alternative has been prevents the loop-control variable from having this value? What will
exercised at least once under some test. happen with this value if it is forced?
 If we do enough tests to achieve this prescription, then we have achieved
100% branch coverage.
 An alternative characterization is to say that we have achieved 100% link CASE 2: Single loop, Non-zero minimum, No excluded values
coverage.

. Dept. of CSE Page 28 Dept. of CSE Page 29

 1. Try one less than the expected minimum. What happens if the loop  If the loops cannot be on the same path, then they are not concatenated
control variable's value is less than the minimum? What prevents the and can be treated as individual loops.
value from being less than the minimum?
2. The minimum number of iterations.
3. One more than the minimum number of iterations. Horrible Loops:
 A horrible loop is a combination of nested loops, the use of code that
4. Once, unless covered by a previous test.
jumps into and out of loops, intersecting loops, hidden loops, and cross
5. Twice, unless covered by a previous test.
connected loops.
6. A typical value.
 Makes iteration value selection for test cases an awesome and ugly task,
7. One less than the maximum value.
which is another reason such structures should be avoided.
8. The maximum number of iterations.
9. Attempt one more than the maximum number of iterations.
CASE 3: Single loops with excluded values
 Treat single loops with excluded values as two sets of tests consisting of
loops without excluded values, such as case 1 and 2 above.
 Example, the total range of the loop control variable was 1 to 20, but that
values 7,8,9,10 were excluded. The two sets of tests are 1-6 and 11-20.
 The test cases to attempt would be 0,1,2,4,6,7 for the first range and
10,11,15,19,20,21 for the second range.
Kinds of Loops:There are only three kinds of loops with respect to path testing:
Nested Loops:
 The number of tests to be performed on nested loops will be the exponent
of the tests performed on single loops.
 As we cannot always afford to test all combinations of nested loops'
iterations values. Here's a tactic used to discard some of these values:
1. Start at the inner most loop. Set all the outer loops to their
minimum values.
2. Test the minimum, minimum+1, typical, maximum-1 , and
maximum for the innermost loop, while holding the outer
loops at their minimum iteration parameter values. Expand the
tests as required for out of range and excluded values.
3. If you've done the outmost loop, GOTO step 5, else move out
one loop and set it up as in step 2 with all other loops set to
typical values.
4. Continue outward in this manner until all loops have been
covered.
5. Do all the cases for all loops in the nest simultaneously.
Concatenated Loops:
 Concatenated loops fall between single and nested loops with respect to
test cases. Two loops are concatenated if it's possible to reach one after
exiting the other while still on a path from entrance to exit.
Figure 2.10: Example of Loop types

Dept. of CSE Page 30 Dept. of CSE Page 31

 Loop Testing TIme: PREDICATE INTERPRETATION:
 Any kind of loop can lead to long testing time, especially if all the extreme value  The simplest predicate depends only on input variables.
cases are to attempted (Max-1, Max, Max+1).  For example if x1,x2 are inputs, the predicate might be x1+x2>=7, given
 This situation is obviously worse for nested and dependent concatenated loops. the values of x1 and x2 the direction taken through the decision is based
 Consider nested loops in which testing the combination of extreme values lead to on the predicate is determined at input time and does not depend on
long test times. Several options to deal with: processing.
 Prove that the combined extreme cases are hypothetically possible,  Another example, assume a predicate x1+y>=0 that along a path prior to
they are not possible in the real world reaching this predicate we had the assignement statement y=x2+7.
 Put in limits or checks that prevent the combined extreme cases. Then although our predicate depends on processing, we can substitute the
you have to test the software that implements such safety measures. symbolic expression for y to obtain an equivalent predicate x1+x2+7>=0.
 The act of symbolic substitution of operations along the path in order to
express the predicate solely in terms of the input vector is
called predicate interpretation.
2. PREDICATES, PATH PREDICATES AND ACHIEVABLE PATHS:  Some times the interpretation may depend on the path; for example,
INPUT X
ON X GOTO A, B, C, ...
A: Z := 7 @ GOTO HEM
PREDICATE: The logical function evaluated at a decision is called Predicate. The B: Z := -7 @ GOTO HEM
direction taken at a decision depends on the value of decision variable. Some examples are: C: Z := 0 @ GOTO HEM
A>0, x+y>=90....... .........
HEM: DO SOMETHING
PATH PREDICATE: A predicate associated with a path is called a Path Predicate. For .........
example, "x is greater than zero", "x+y>=90", "w is either negative or equal to 10 is true" is HEN: IF Y + Z > 0 GOTO ELL ELSE GOTO EMM
a sequence of predicates whose truth values will cause the routine to take a specific path. The predicate interpretation at HEN depends on the path we took
through the first multiway branch. It yields for the three cases
MULTIWAY BRANCHES: respectively, if Y+7>0, Y-7>0, Y>0.
 The path taken through a multiway branch such as a computed GOTO's,
The path predicates are the specific form of the predicates of the decisions
case statement, or jump tables cannot be directly expressed in along the selected path after interpretation.
TRUE/FALSE terms.
 Although, it is possible to describe such alternatives by using multi valued INDEPENDENCE OF VARIABLES AND PREDICATES:
logic, an expedient (practical approach) is to express multiway branches
as an equivalent set of if..then..else statements.  The path predicates take on truth values based on the values of input
 For example a three way case statement can be written as: If case=1 DO variables, either directly or indirectly.
A1 ELSE (IF Case=2 DO A2 ELSE DO A3 ENDIF)ENDIF.  If a variable's value does not change as a result of processing, that variable
INPUTS: is independent of the processing.
 In testing, the word input is not restricted to direct inputs, such as  If the variable's value can change as a result of the processing, the variable
variables in a subroutine call, but includes all data objects referenced by is process dependent.
the routine whose values are fixed prior to entering it.  A predicate whose truth value can change as a result of the processing is
 For example, inputs in a calling sequence, objects in a data structure, said to be process dependent and one whose truth value does not change
values left in registers, or any combination of object types. as a result of the processing is process independent.
 The input for a particular test is mapped as a one dimensional array called  Process dependence of a predicate does not always follow from
as an Input Vector. dependence of the input variables on which that predicate is based.

Dept. of CSE Page 32 Dept. of CSE Page 33

CORRELATION OF VARIABLES AND PREDICATES: 
As achieving the desired direction at a given decision could still hide bugs
 Two variables are correlated if every combination of their values cannot in the associated predicates.
be independently specified. TESTING BLINDNESS:
 Variables whose values can be specified independently without restriction  Testing Blindness is a pathological (harmful) situation in which the desired path is
are called uncorrelated. achieved for the wrong reason.
 A pair of predicates whose outcomes depend on one or more variables in  There are three types of Testing Blindness:
common are said to be correlated predicates. 0. Assignment Blindness:
For example, the predicate X==Y is followed by another predicate X+Y  Assignment blindness occurs when the buggy predicate appears to
== 8. If we select X and Y values to satisfy the first predicate, we might work correctly because the specific value chosen for an assignment
have forced the 2nd predicate's truth value to change. statement works with both the correct and incorrect predicate.
 Every path through a routine is achievable only if all the predicates in that  For Example:
routine are uncorrelated. Correct Buggy
PATH PREDICATE EXPRESSIONS:
 A path predicate expression is a set of boolean expressions, all of which
X=7 X=7
........ ........
must be satisfied to achieve the selected path.
if Y > 0 then ... if X+Y > 0 then ...
 Example:
 If the test case sets Y=1 the desired path is taken in either case, but
X1+3X2+17>=0
X3=17 there is still a bug.
X4-X1>=14X2 2. Equality Blindness:
 Equality blindness occurs when the path selected by a prior predicate
 Any set of input values that satisfy all of the conditions of the path results in a value that works both for the correct and buggy predicate.
predicate expression will force the routine to the path.  For Example:
 Some times a predicate can have an OR in it. Correct Buggy
 Example:
if Y = 2 then if Y = 2 then
A: X5 > 0 E: X6 < 0 ........ ........
B: X1 + 3X2 + 17 >= 0 B: X1 + 3X2 + 17 >= 0 if X+Y > 3 then ... if X > 1 then ...
C: X3 = 17 C: X3 = 17
D: X4 - X1 >= 14X2 D: X4 - X1 >= 14X2  The first predicate if y=2 forces the rest of the path, so that for any
positive value of x. the path taken at the second predicate will be the

Boolean algebra notation to denote the boolean expression: same for the correct and buggy version.
ABCD+EBCD=(A+E)BCD 3. Self Blindness:
PREDICATE COVERAGE:  Self blindness occurs when the buggy predicate is a multiple of the
 Compound Predicate: Predicates of the form A OR B, A AND B and correct predicate and as a result is indistinguishable along that path.
more complicated boolean expressions are called as compound  For Example:
predicates. Correct Buggy
 Some times even a simple predicate becomes compound after
X =A X =A
interpretation. Example: the predicate if (x=17) whose opposite branch is
........ ........
if x.NE.17 which is equivalent to x>17 . Or. X<17. if X-1 > 0 then ... if X+A-2 > 0 then ...
 Predicate coverage is being the achieving of all possible combinations of
 The assignment (x=a) makes the predicates multiples of each other,
truth values corresponding to the selected path have been explored under
so the direction taken is the same for the correct and buggy version.
some test.

. Dept. of CSE Page 34 Dept. of CSE Page 35

3. PATH SENSITIZING:
4. PATH INSTRUMENTATION
REVIEW: ACHIEVABLE AND UNACHIEVABLE PATHS:
 We want to select and test enough paths to achieve a satisfactory notion of test PATH INSTRUMENTATION:
completeness such as C1+C2.  Path instrumentation is what we have to do to confirm that the outcome was achieved
 Extract the programs control flowgraph and select a set of tentative covering paths. by the intended path.
 For any path in that set, interpret the predicates along the path as needed to express  Co-incidental Correctness: The coincidental correctness stands for achieving the
them in terms of the input vector. In general individual predicates are compound or desired outcome for wrong reason.
may become compound as a result of interpretation.
 Trace the path through, multiplying the individual compound predicates to achieve
a boolean expression such as
(A+BC) (D+E) (FGH) (IJ) (K) (l) (L).
 Multiply out the expression to achieve a sum of products form:
ADFGHIJKL+AEFGHIJKL+BCDFGHIJKL+BCEFGHIJKL
 Each product term denotes a set of inequalities that if solved will yield an input
vector that will drive the routine along the designated path.
 Solve any one of the inequality sets for the chosen path and you have found a set of
input values for the path.
 If you can find a solution, then the path is achievable.
 If you cant find a solution to any of the sets of inequalities, the path is un
achievable.
 The act of finding a set of solutions to the path predicate expression is called PATH
SENSITIZATION. Figure 1.11: Coincidental Correctness
The above figure is an example of a routine that, for the (unfortunately) chosen input value (X =
HEURISTIC PROCEDURES FOR SENSITIZING PATHS: 16), yields the same outcome (Y = 2) no matter which case we select. Therefore, the tests chosen
 This is a workable approach, instead of selecting the paths without considering this way will not tell us whether we have achieved coverage. For example, the five cases could be
how to sensitize, attempt to choose a covering path set that is easy to sensitize and totally jumbled and still the outcome would be the same. Path Instrumentation is what we have
pick hard to sensitize paths only as you must to achieve coverage. to do to confirm that the outcome was achieved by the intended path.
 Identify all variables that affect the decision.
 Classify the predicates as dependent or independent.
The types of instrumentation methods include:
1. Interpretive Trace Program:
 Start the path selection with un correlated, independent predicates.
 An interpretive trace program is one that executes every statement in order and
 If coverage has not been achieved using independent uncorrelated predicates,
records the intermediate values of all calculations, the statement labels
extend the path set using correlated predicates.
traversed etc.
 If coverage has not been achieved extend the cases to those that involve dependent
 If we run the tested routine under a trace, then we have all the information we
predicates.
need to confirm the outcome and, furthermore, to confirm that it was achieved
 Last, use correlated, dependent predicates.
by the intended path.
 The trouble with traces is that they give us far more information than we need.
In fact, the typical trace program provides so much information that confirming

Dept. of CSE Page 36 Dept. of CSE Page 37

 the path from its massive output dump is more work than simulating the
computer by hand to confirm the path.
2. Traversal Marker or Link Marker:
 A simple and effective form of instrumentation is called a traversal marker or
link marker.
 Name every link by a lower case letter.
 Instrument the links so that the link's name is recorded when the link is
executed.
 The succession of letters produced in going from the routine's entry to its exit
should, if there are no bugs, exactly correspond to the path name.

Figure 2.14: Double Link Marker Instrumentation.

4. Link Counter: A less disruptive (and less informative) instrumentation method is
based on counters. Instead of a unique link name to be pushed into a string when the link
is traversed, we simply increment a link counter. We now confirm that the path length is
as expected. The same problem that led us to double link markers also leads us to double
link counters.

Figure 1.12: Single Link Marker Instrumentation

 Why Single Link Markers aren't enough: Unfortunately, a single link
marker may not do the trick because links can be chewed by open bugs.

Figure 2.13: Why Single Link Markers aren't enough.

We intended to traverse the ikm path, but because of a rampaging GOTO in the
middle of the m link, we go to process B. If coincidental correctness is against us, the outcomes
will be the same and we won't know about the bug.
3. Two Link Marker Method:
 The solution to the problem of single link marker method is to implement two
markers per link: one at the beginning of each link and on at the end.
 The two link markers now specify the path name and confirm both the
beginning and end of the link.

Dept. of CSE Page 38 Dept. of CSE Page 39