Testing Phases

• Coverage
– ideally, testing will exercise the system in all
possible ways
– not possible, so we use different criteria to judge
how well our testing strategy “covers” the system
• Test case
– consists of data, procedure, and expected result
– represents just one situation under which the system
(or some part of it) might run
Testing Phases
In development/
In user
• Unit testing - does this piece work by
• Integration testing - do these two
pieces work together?
• System testing - do all the pieces work
• Alpha acceptance testing - try it out
with in-house “users”
• Installation testing - can users install
it and does it work in their
• Beta acceptance testing - try it out
with real users
Test Planning
• A Test Plan:
covers all types and phases of testing
guides the entire testing process
who, why, when, what
developed as requirements, functional
specification, and high-level design are
– should be done before implementation starts
Test Planning (cont.)
• A test plan includes:
test objectives
schedule and logistics
test strategies
test cases
• procedure
• data
• expected result
– procedures for handling problems
Testing Techniques
• Structural testing techniques
– “white box” testing
– based on statements in the code
– coverage criteria related to physical parts of the
– tests how a program/system does something
• Functional testing techniques
“black box” testing
based on input and output
coverage criteria based on behavior aspects
tests the behavior of a system or program
System Testing Techniques
• Goal is to evaluate the system as a whole,
not its parts
• Techniques can be structural or functional
• Techniques can be used in any stage that
tests the system as a whole (acceptance,
installation, etc.)
• Techniques not mutually exclusive
System Testing Techniques (cont.)
• Structural techniques
– stress testing - test larger-than-normal capacity
in terms of transactions, data, users, speed, etc.
– execution testing - test performance in terms of
speed, precision, etc.
– recovery testing - test how the system recovers
from a disaster, how it handles corrupted data,
System Testing Techniques (cont.)
• Structural techniques (cont.)
– operations testing - test how the system fits in
with existing operations and procedures in the
user organization
– compliance testing - test adherence to standards
– security testing - test security requirements
System Testing Techniques (cont.)
• Functional techniques
– requirements testing - fundamental form of testing makes sure the system does what it’s required to do
– regression testing - make sure unchanged
functionality remains unchanged
– error-handling testing - test required error-handling
functions (usually user error)
– manual-support testing - test that the system can be
used properly - includes user documentation
System Testing Techniques (cont.)
• Functional techniques (cont.)
– intersystem handling testing - test that the
system is compatible with other systems in the
– control testing - test required control
– parallel testing - feed same input into two
versions of the system to make sure they
produce the same output
Unit Testing
• Goal is to evaluate some piece (file,
program, module, component, etc.) in
• Techniques can be structural or functional
• In practice, it’s usually ad-hoc and looks a
lot like debugging
• More structured approaches exist
Unit Testing Techniques
• Functional techniques
– input domain testing - pick test cases
representative of the range of allowable input,
including high, low, and average values
– equivalence partitioning - partition the range of
allowable input so that the program is expected
to behave similarly for all inputs in a given
partition, then pick a test case from each
Unit Testing Techniques (cont.)
• Functional techniques (cont.)
– boundary value - choose test cases with input values
at the boundary (both inside and outside) of the
allowable range
– syntax checking - choose test cases that violate the
format rules for input
– special values - design test cases that use input
values that represent special situations
– output domain testing - pick test cases that will
produce output at the extremes of the output domain
Unit Testing Techniques (cont.)
• Structural techniques
– statement testing - ensure the set of test cases
exercises every statement at least once
– branch testing - each branch of an if/then
statement is exercised
– conditional testing - each truth statement is
exercised both true and false
– expression testing - every part of every
expression is exercised
– path testing - every path is exercised
(impossible in practice)
Unit Testing Techniques (cont.)
• Error-based techniques
– basic idea is that if you know something about
the nature of the defects in the code, you can
estimate whether or not you’ve found all of
them or not
– fault seeding - put a certain number of known
faults into the code, then test until they are all
Unit Testing Techniques (cont.)
• Error-based techniques (cont.)
– mutation testing - create mutants of the
program by making single changes, then run
test cases until all mutants have been killed
– historical test data - an organization keeps
records of the average numbers of defects in the
products it produces, then tests a new product
until the number of defects found approaches
the expected number
Testing vs. Static Analysis
• Static Analysis attempts to evaluate a program
without executing it
• Techniques:
inspections and reviews
mathematical correctness proofs
safety analysis
program measurements
data flow and control flow analysis
symbolic execution
Testing: A Roadmap
by Mary Jean Harrold, Future of Software Engineering
Series, ICSE 2000, Limerick, Ireland
Status of
Tools, and
Processes in
Fundamental Research
Testing componentbased systems
Testing based on
precode artifacts
Testing evolving
Demonstrating effectiveness
of testing techniques
Creating effective
testing processes
Using testing
Other testing
Practical Testing
Methods, Tools,
and Processes for
Development of
Testing: A Roadmap (cont.)
• Regression testing can account for up to 1/3 of
total cost of software
• Research has focused on ways to make regression
testing more efficient:
– Finding most effective subsets of the test suite
– Managing the growth of the test suite
– Use precode artifacts to help plan regression testing
• Need to focus on reducing size of test suite,
prioritizing test cases, and assessing testability
Test Case Prioritization: An Empirical Study
Gregg Rothermel et al, ICSM 1999, Oxford, England
• Prioritization is necessary:
– during regression testing
– when there is not enough time or resources to run the
entire test suite
– when there is a need to decide which test cases to run
first, i.e. which give the biggest bang for the buck
• Prioritization can be based on:
– which test cases are likely to find the most faults
– which test cases are likely to find the most severe faults
– which test cases are likely to cover the most statements
Test Case Prioritization: An
Empirical Study
• The study:
– Applied 9 different prioritization techniques to
7 different programs
– All prioritization techniques were aimed at
finding faults faster
– Results showed that prioritization definitely
increases the fault detection rate, but different
prioritization techniques didn’t differ much in
A Manager’s Guide to Evaluating Test Suites
Brian Marick, et al. http://www.testing.com/writings/evaluating-testsuites-paper.pdf
• Basic assumption: purpose of testing is to
help manager decide whether or not to
release the system to the customer
• Threat Product Problem Victim
• Users - create detailed descriptions of
fictitious, but realistic, users
• Pay attention to usability
• Coverage - code, documentation
Summary of Readings
• Rigorous scientific study vs. practical
– Harrold and Rothermel are academic
– Marick is a practitioner/consultant
– differences in perspective, concerns,
• Plan ahead vs. evaluate after the fact
Summary of Readings (cont.)
• Regression testing
big cost
size and quality of test suite is crucial
automation only goes so far
tradeoff between test planning time and test
execution time