Transcript Verification and Validation

Verification and Validation
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 1
Objectives




To introduce software verification and validation and
to discuss the distinction between them
To describe the program inspection process and its
role in V & V
To explain static analysis as a verification technique
To describe the Cleanroom software development
process
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 2
Topics covered







Introduction to testing
Sources of errors
Why do we need testing ?
Verification and validation planning
Software inspections
Automated static analysis
Cleanroom software development
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 3
Nasty question



Suppose you are being asked to lead the team to
test the software that controls a new X-ray machine.
Would you take that job?
Would you take it if you could name your own price?
What if the contract says you’ll be charged with
murder in case a patient dies because of a malfunctioning of the software?
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 4
A few spectacular software failures
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 5
Failures in Production Software





NASA’s Mars lander, September 1999, crashed due to a units integration
fault—over $50 million US !
Huge losses due to web application failures
•
Financial services : $6.5 million per hour
•
Credit card sales applications : $2.4 million per hour
In Dec 2006, amazon.com’s BOGO offer turned into a double discount
2007 : Symantec says that most security vulnerabilities are due to faulty
software
Stronger testing could solve most of these problems
World-wide monetary loss due to poor software is staggering
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
6
Thanks to Dr. Sreedevi Sampath
Slide 6
Bypass Testing Results
v
— Vasileios Papadimitriou. Masters thesis, Automating Bypass Testing for Web Applications, GMU 2006
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 7
Why Does Testing Matter?




Ariane 5:
exception-handling
NIST report, “The Economic Impacts of Inadequate bug : forced self
Infrastructure for Software Testing” (2002)
– Inadequate software testing costs the US alone destruct on maiden
flight (64-bit to 16-bi
between $22 and $59 billion annually
– Better approaches could cut this amount in half conversion: about
Major failures: Ariane 5 explosion, Mars Polar
370 million $ lost)
Lander, Intel’s Pentium FDIV bug
Insufficient testing of safety-critical software can cost
lives:
 THERAC-25 radiation machine: 3 dead
We want our programs to be reliable
– Testing is how, in most cases, we find out if they
are
Sommerville, Ammann-Offutt, Mejia, 2009
THERAC-25
Software
Engineering
design
Mars Polar
Lander
crash
site?
8
Slide 8
Software is a Skin that Surrounds Our
Civilization
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 9
Airbus 319 Safety Critical
Software Control
Loss of autopilot
Loss of most flight deck lighting and intercom
Loss of both the commander’s and the
co-pilot’s primary flight and navigation displays
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 10
Northeast Blackout of 2003
508 generating
units and 256
power plants shut
down
Affected 10 million
people in Ontario,
Canada
Affected 40 million
people in 8 US
states
Financial losses of
$6 Billion USD
The alarm system in the energy management system failed
due to a software error and operators were not informed of
the power overload in the system
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 11
Software testing is getting more
important
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 12
Testing in the 21st Century





We are going through a time of change
Software defines behavior
is going
•
network routers, finance, switching networks,Industry
other infrastructure
through a revolution in
Today’s software market :
•
is much bigger
what testing means to
•
is more competitive
the success of software
•
has more users
products
Agile processes put increased pressure on testers
Embedded Control Applications
•
airplanes, air traffic control
– PDAs
•
spaceships
– memory seats
•
watches
– DVD players
•
ovens
– garage door openers
•
remote controllers
– cell phones
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 13
Testing in the 21st Century






More safety critical, real-time software
Enterprise applications means bigger programs, more users
Embedded software is ubiquitous … check your pockets
Paradoxically, free software increases our expectations !
Security is now all about software faults
•
Secure software is reliable software
The web offers a new deployment platform
•
Very competitive and very available to more users
•
Web apps are distributed
•
Web apps must be highly reliable
Industry desperately needs our inventions !
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 14
Mismatch in Needs and Goals




Industry wants testing to be simple and easy
•
Testers with no background in computing or math
Universities are graduating scientists
•
Industry needs engineers
Testing needs to be done more rigorously
Agile processes put lots of demands on testing
•
Programmers must unit test – with no training, education
or tools !
•
Tests are key components of functional requirements –
but who builds those tests ?
Bottom line—lots of crappy software
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 15
Here! Test This!
My first “professional” job
Big
software program
MicroSteff
– big
software system
for the mac
V.1.5.1
Jan/2007
Jan/2007
MF2-HD
1.44 MB
DataLife
Verdatim
A stack of computer printouts—and no documentation
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 16
Cost of Testing
You’re going to spend at least half of
your development budget on testing,
whether you want to or not



In the real-world, testing is the principle post-design
activity
Restricting early testing usually increases cost
Extensive hardware-software integration requires
more testing
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
17
Slide 17
State-of-the-Art




30-85 errors are made per 1000 lines of source code
extensively tested software contains 0.5-3 errors per
1000 lines of source code
testing is postponed, as a consequence: the later an
error is discovered, the more it costs to fix it.
error distribution: 60% design, 40% implementation.
66% of the design errors are not discovered until the
software has become operational.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 18
Relative cost of error correction
100
50
20
10
5
2
1
RE
design
Sommerville, Ammann-Offutt, Mejia, 2009
code
Software Engineering
test
operation
Slide 19
Why Test?
If you don’t know why you’re conducting
a test, it won’t be very helpful

Written test objectives and requirements are rare

What are your planned coverage levels?

How much testing is enough?

Common objective – spend the budget …
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 20
Lessons

Many errors are made in the early phases

These errors are discovered late

Repairing those errors is costly

 It pays off to start testing real early
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 21
Why Test?
If you don’t start planning for each test when
the functional requirements are formed, you’ll
never know why you’re conducting the test

1980: “The software shall be easily maintainable”

Threshold reliability requirements?

What fact is each test trying to verify?

Requirements definition teams should include testers!
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
22
Slide 22
Cost of Not Testing
Program Managers often say:
“Testing is too expensive.”




Not testing is even more expensive
Planning for testing after development is prohibitively
expensive
A test station for circuit boards costs half a million dollars
…
Software test tools cost less than $10,000 !!!
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 23
How then to proceed?



Exhaustive testing most often is not feasible
Random statistical testing does not work
either if you want to find errors
Therefore, we look for systematic ways to
proceed during testing
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 24
Some preliminary questions

What exactly is an error?

How does the testing process look like?

When is test technique A superior to test technique
B?

What do we want to achieve during testing?

When to stop testing?
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 25
Error, fault, failure

an error is a human activity resulting in
software containing a fault

a fault is the manifestation of an error

a fault may result in a failure
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 26
When exactly is a failure a
failure?



Failure is a relative notion: e.g. a failure w.r.t. the
specification document
Verification: evaluate a product to see whether it
satisfies the conditions specified at the start:
Have we built the system right?
Validation: evaluate a product to see whether it does
what we think it should do:
Have we built the right system?
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 27
What is our goal during testing?


Objective 1: find as many faults as possible
Objective 2: make you feel confident that the
software works OK
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 28
Verification vs validation




Verification:
"Are we building the product right”.
The software should conform to its
specification.
Validation:
"Are we building the right product”.
The software should do what the user really
requires.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 29
The V & V process


Is a whole life-cycle process - V & V must be
applied at each stage in the software
process.
Has two principal objectives
•
•
The discovery of defects in a system;
The assessment of whether or not the system is
useful and useable in an operational situation.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 30
V& V goals



Verification and validation should establish
confidence that the software is fit for
purpose.
This does NOT mean completely free of
defects.
Rather, it must be good enough for its
intended use and the type of use will
determine the degree of confidence that is
needed.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 31
V & V confidence

Depends on system’s purpose, user
expectations and marketing environment
•
Software function
• The level of confidence depends on how critical the
software is to an organisation.
•
User expectations
• Users may have low expectations of certain kinds of
software.
•
Marketing environment
• Getting a product to market early may be more
important than finding defects in the program.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 32
Static and dynamic verification

Software inspections. Concerned with analysis of
the static system representation to discover
problems (static verification)
•

May be supplement by tool-based document and code
analysis
Software testing. Concerned with exercising and
observing product behaviour (dynamic verification)
•
The system is executed with test data and its operational
behaviour is observed
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 33
Static and dynamic V&V
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 34
Program testing



Can reveal the presence of errors NOT their
absence.
The only validation technique for nonfunctional requirements as the software has
to be executed to see how it behaves.
Should be used in conjunction with static
verification to provide full V&V coverage.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 35
Types of testing

Defect testing
•
•
•

Tests designed to discover system defects.
A successful defect test is one which reveals the
presence of defects in a system.
Covered in Chapter 23
Validation testing
•
•
Intended to show that the software meets its
requirements.
A successful test is one that shows that a requirements
has been properly implemented.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 36
Testing and debugging




Defect testing and debugging are distinct
processes.
Verification and validation is concerned with
establishing the existence of defects in a program.
Debugging is concerned with locating and
repairing these errors.
Debugging involves formulating a hypothesis
about program behaviour then testing these
hypotheses to find the system error.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 37
The debugging process
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 38
V & V planning




Careful planning is required to get the most
out of testing and inspection processes.
Planning should start early in the
development process.
The plan should identify the balance
between static verification and testing.
Test planning is about defining standards for
the testing process rather than describing
product tests.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 39
The V-model of development
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 40
The structure of a software test plan







The testing process.
Requirements traceability.
Tested items.
Testing schedule.
Test recording procedures.
Hardware and software requirements.
Constraints.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 41
The software test plan
The testing process
A description of the major phases of the testing process. These might be
as described earlier in t his chapter.
Requirements traceability
Users are most interested in the system meeting its requirements and
testing should be planned so that all requirements are individually tested.
Tested items
The products of the software process that are to be tested should be
specified.
Testing schedule
An overall testing schedule and resource allocation for this schedule.
This, obvi ously, is linked to the more general project development
schedule.
Test recording procedures
It is not enough simply to run tests. The results of the tests must be
systematically recorded. It must be possible to audit the testing process
to check that it been carried out correctly.
Hardwa re and software requirements
This section should set out software tools required and estimated
hardware utilisation.
Constraints
Constraints affecting the testing process such as staff shortages should
be anticipated in this section.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 42
Software inspections




These involve people examining the source
representation with the aim of discovering anomalies
and defects.
Inspections not require execution of a system so
may be used before implementation.
They may be applied to any representation of the
system (requirements, design,configuration data,
test data, etc.).
They have been shown to be an effective technique
for discovering program errors.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 43
Inspection success


Many different defects may be discovered in
a single inspection. In testing, one defect
,may mask another so several executions
are required.
The reuse domain and programming
knowledge so reviewers are likely to have
seen the types of error that commonly arise.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 44
Inspections and testing




Inspections and testing are complementary and not
opposing verification techniques.
Both should be used during the V & V process.
Inspections can check conformance with a
specification but not conformance with the
customer’s real requirements.
Inspections cannot check non-functional
characteristics such as performance, usability, etc.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 45
Program inspections



Formalised approach to document reviews
Intended explicitly for defect detection (not
correction).
Defects may be logical errors, anomalies in
the code that might indicate an erroneous
condition (e.g. an uninitialised variable) or
non-compliance with standards.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 46
Inspection pre-conditions






A precise specification must be available.
Team members must be familiar with the
organisation standards.
Syntactically correct code or other system
representations must be available.
An error checklist should be prepared.
Management must accept that inspection will
increase costs early in the software process.
Management should not use inspections for staff
appraisal ie finding out who makes mistakes.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 47
The inspection process
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 48
Inspection procedure





System overview presented to inspection
team.
Code and associated documents are
distributed to inspection team in advance.
Inspection takes place and discovered errors
are noted.
Modifications are made to repair discovered
errors.
Re-inspection may or may not be required.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 49
Inspection roles
Author or owner
The programmer or designer responsible fo r
producing the program or document. Responsible
for fixing defects discovered during the inspection
process.
Inspector
Finds errors, omissions and inconsistencies in
programs and documents. May also identify
broader issues that are outside the scope of the
inspection team.
Reader
Presents the code or document at an inspection
meeting.
Scribe
Records the results of the inspection meeting.
Chairman or mo derator
Manages the process and facilitates the inspection.
Reports process results to the Chief mo derator.
Chief mo derator
Responsible for inspection process imp rovements,
checklist updating, standards development etc.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 50
Inspection checklists




Checklist of common errors should be used to
drive the inspection.
Error checklists are programming language
dependent and reflect the characteristic errors that
are likely to arise in the language.
In general, the 'weaker' the type checking, the larger
the checklist.
Examples: Initialisation, Constant naming, loop
termination, array bounds, etc.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 51
Inspection checks 1
Data faults
Are all program variables initialised before their values are
used?
Have all constants been named?
Should the upper bound of arrays be equal to the size of the
array or Size -1?
If character strings are used, is a de limiter explicitly
assigned?
Is there any possibility of b uffer overflow?
Control faults
For each conditional statement, is the condition correct?
Is each loop certain to terminate?
Are comp ound statements correctly bracketed?
In case statements, are all possible cases accounted for?
If a break is required after each case in case statements, has
it been included?
Input/output faults
Are all input variables used?
Are all output variables assigned a value before they are
output?
Can unexpected inputs cause corruption?
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 52
Inspection checks 2
Interface faults
Do all function and method calls have the correct number
of parameters?
Do fo rmal and actual parameter types match?
Are the parameters in the right order?
If comp onents access shared memo ry, do they have the
same mo del of the shared memo ry structure?
Storage
If a linked structure is modified, have all links been
manageme nt faults correctly reassigned?
If dynami c storage is used, has space been allocated
correctly?
Is space explicitly de-allocated after it is no longer
required?
Exception
Have all possible error conditions been taken into account?
manageme nt faults
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 53
Inspection rate





500 statements/hour during overview.
125 source statement/hour during individual
preparation.
90-125 statements/hour can be inspected.
Inspection is therefore an expensive
process.
Inspecting 500 lines costs about 40
man/hours effort - about £2800 at UK rates.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 54
Automated static analysis



Static analysers are software tools for source
text processing.
They parse the program text and try to
discover potentially erroneous conditions and
bring these to the attention of the V & V
team.
They are very effective as an aid to
inspections - they are a supplement to but
not a replacement for inspections.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 55
Static analysis checks
Fault class
Static analysis check
Data faults
Variables used befo re initialisation
Variables declared but never used
Variables assigned twice but never used between
assignments
Possible array bound violations
Undeclared variables
Control faults
Unreachable code
Unconditional branches into loops
Input/output faults
Variables output twice with no intervening
assignment
Interface faults
Parameter type mismatches
Parameter numb er mismatches
Non-usage of t he results of f unctions
Uncalled functions and procedures
Storage ma nagement
faults
Unassigned pointers
Pointer arithmetic
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 56
Stages of static analysis



Control flow analysis. Checks for loops with
multiple exit or entry points, finds unreachable
code, etc.
Data use analysis. Detects uninitialised
variables, variables written twice without an
intervening assignment, variables which are
declared but never used, etc.
Interface analysis. Checks the consistency of
routine and procedure declarations and their
use
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 57
Stages of static analysis



Information flow analysis. Identifies the
dependencies of output variables. Does not
detect anomalies itself but highlights
information for code inspection or review
Path analysis. Identifies paths through the
program and sets out the statements
executed in that path. Again, potentially
useful in the review process
Both these stages generate vast amounts of
information. They must be used with care.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 58
LINT static analysis
138% more lint_ex.c
#include <stdio.h>
printarray (Anarray)
int A narray;
{ printf(“%d”,Anarray); }
main ()
{
int A narray[ 5]; int i; char c;
printarray (Anarray, i, c);
printarray (Anarray) ;
}
139% cc lint_ex.c
140% lint lint_ex.c
lint_ex.c(10): warning: c may b e used before set
lint_ex.c(10): warning: i may be used before set
printarray: variable # of args. lint_ex.c(4) :: lint_ex.c(10)
printarray, arg. 1 used inconsistentlyli nt_ex.c(4) :: lint_ex.c(10)
printarray, arg. 1 used inconsistentlyli nt_ex.c(4) :: lint_ex.c(11)
printf returns value which is always ignored
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 59
Use of static analysis


Particularly valuable when a language such
as C is used which has weak typing and
hence many errors are undetected by the
compiler,
Less cost-effective for languages like Java
that have strong type checking and can
therefore detect many errors during
compilation.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 60
Verification and formal methods



Formal methods can be used when a
mathematical specification of the system is
produced.
They are the ultimate static verification
technique.
They involve detailed mathematical analysis
of the specification and may develop formal
arguments that a program conforms to its
mathematical specification.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 61
Arguments for formal methods


Producing a mathematical specification
requires a detailed analysis of the
requirements and this is likely to uncover
errors.
They can detect implementation errors
before testing when the program is analysed
alongside the specification.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 62
Arguments against formal methods



Require specialised notations that cannot be
understood by domain experts.
Very expensive to develop a specification
and even more expensive to show that a
program meets that specification.
It may be possible to reach the same level of
confidence in a program more cheaply using
other V & V techniques.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 63
Cleanroom software development


The name is derived from the 'Cleanroom'
process in semiconductor fabrication. The
philosophy is defect avoidance rather than
defect removal.
This software development process is based on:
•
•
•
•
Incremental development;
Formal specification;
Static verification using correctness arguments;
Statistical testing to determine program reliability.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 64
The Cleanroom process
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 65
Cleanroom process characteristics





Formal specification using a state transition
model.
Incremental development where the
customer prioritises increments.
Structured programming - limited control and
abstraction constructs are used in the
program.
Static verification using rigorous inspections.
Statistical testing of the system (covered in
Ch. 24).
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 66
Formal specification and inspections



The state based model is a system
specification and the inspection process
checks the program against this mode.l
The programming approach is defined so
that the correspondence between the model
and the system is clear.
Mathematical arguments (not proofs) are
used to increase confidence in the inspection
process.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 67
Cleanroom process teams



Specification team. Responsible for developing
and maintaining the system specification.
Development team. Responsible for
developing and verifying the software. The
software is NOT executed or even compiled
during this process.
Certification team. Responsible for developing
a set of statistical tests to exercise the software
after development. Reliability growth models
used to determine when reliability is acceptable.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 68
Cleanroom process evaluation




The results of using the Cleanroom process have
been very impressive with few discovered faults in
delivered systems.
Independent assessment shows that the
process is no more expensive than other
approaches.
There were fewer errors than in a 'traditional'
development process.
However, the process is not widely used. It is not
clear how this approach can be transferred
to an environment with less skilled or less
motivated software engineers.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 69
Key points



Verification and validation are not the same
thing. Verification shows conformance with
specification; validation shows that the
program meets the customer’s needs.
Test plans should be drawn up to guide the
testing process.
Static verification techniques involve
examination and analysis of the program for
error detection.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 70
Key points




Program inspections are very effective in
discovering errors.
Program code in inspections is systematically
checked by a small team to locate software faults.
Static analysis tools can discover program
anomalies which may be an indication of faults in the
code.
The Cleanroom development process depends on
incremental development, static verification and
statistical testing.
Sommerville, Ammann-Offutt, Mejia, 2009
Software Engineering
Slide 71