cs498dm software testing darko marinov january 17, 2012

cs498dmSoftware Testing

Darko Marinov

January 17, 2012

Teaching Staff

• Instructor: Darko Marinov– Email: marinov AT illinois.edu– Office: SC 3116, 217-265-6117– Office hours: after classes or by appointment

• No TA

Course Overview

• Introduction to software testing– Systematic, organized approaches to testing– Based on models and coverage criteria– Testing is not (only) about finding “bugs”– Improve your testing (and development) skills

• Five problem sets and a project– Centered around testing Java PathFinder (JPF)

JPF in One Slide

• Will have an entire lecture on JPF– You’ll get to learn a lot about Java/JVM/JPF in

this class

• JPF is a tool for systematic testing of Java programs

• JPF is a Java Virtual Machine (JVM) that has support for fast backtracking

• JPF is implemented in Java itself

Administrative Info

• Lectures: TR 12:30pm-1:45pm, 1103 SC

• Credit:– 3 undergraduate hours– 3 or 4 graduate hours (a larger project for 4)

• Prerequisites: recommended software engineering (cs427) and programming languages (cs225, cs421)– Consent of instructor (must if not senior)

Grading

• Points– Project (25%)– Problem sets (5*15%)

• Grades– A*(90%), B*(80%), C*(70%), D*(60%), F(<60%)– For more details, see the syllabus– The instructor may lower the point limits

Project

• Testing a part of JPF

• Deliverables– Proposal (due in three weeks)– Progress report (around midterm)– Final report (by the grade submission deadline)– Bug reports (hopefully you’ll find some bugs)

• Extra bonus points for reporting bugs to me

Collaboration

• You must individually write solutions for the problem sets

• You can collaborate on everything else (unless explicitly stated not to collaborate!)– Discuss problem sets– Do projects in groups, preferably two or three

students

• Testing is a social activity– Communication matters

Course Communication

• Wikihttps://wiki.engr.illinois.edu/display/cs498dmsp12

• Mailing list cs498dm AT cs.illinois.edu

• Instructor: Darko– Email: marinov AT illinois.edu– Office: SC 3116, 217-265-6117– Office hours: after classes or by appointment

Signup Sheet

• Name

• Email address (NetID@uiuc.edu)

• Program/Year

• Interests: what would you like to learn about testing?

• Experience: what testing did you do?

Textbook

• “Introduction to Software Testing”by Paul Ammann and Jeff OffuttCambridge University Press, Jan. 2008

• Strongly recommended but not required– Books should be in the bookstore already

This Lecture: Introduction to “Bugs”

• Why look for bugs?

• What are bugs?

• Where they come from?

• How to detect them?

Some Costly “Bugs”

• NASA Mars space missions– Priority inversion (2004)– Different metric systems (1999)

• BMW airbag problems (1999)– Recall of 15,000+ cars

• Ariane 5 crash (1996)– Uncaught exception of numerical overflow– http://www.youtube.com/watch?v=kYUrqdUyEpI

• Your own favorite examples?

Some “Bugging” Bugs

• Smaller issues that give unexpected results

• Your own favorite examples?

Economic Impact

• “The Economic Impact of Inadequate Infrastructure for Software Testing”NIST Report, May 2002

• $59.5B annual cost of inadequate software testing infrastructure

• $22.2B annual potential cost reduction from feasible infrastructure improvements

Estimates

• Extrapolated from two studies (5% of total)– Manufacturing: transportation equipment– Services: financial institutions

• Number of simplifying assumptions

• “…should be considered approximations”

• What is important to you?– Correctness, performance, functionality

Some Motivation for Testers

• An article from SD Times, a magazine for software development managers:“Improving Software Quality”by Lindsey Vereen (page 34/68)

• A slide from Debra Richardson, a professor at UC Irvine:“Analysis and Testing are Creative”(page 26/48)

Terminology• Anomaly• Bug• Crash• Defect• Error• Failure, fault • Glitch• Hang• Incorrectness• J...

Dynamic vs. Static

• Incorrect (observed) behavior– Failure, fault

• Incorrect (unobserved) state– Error, latent error

• Incorrect lines of code– Fault, error

“Bugs” in IEEE 610.12-1990

• Fault– Incorrect lines of code

• Error– Faults cause incorrect (unobserved) state

• Failure– Errors cause incorrect (observed) behavior

• Not used consistently in literature!

Correctness and Quality

Common (partial) propertiesSegfaults, uncaught exceptionsResource leaksData races, deadlocksStatistics based

Specific propertiesRequirementsSpecification

Traditional Waterfall ModelRequirements

Analysis

DesignChecking

ImplementationUnit Testing

IntegrationSystem Testing

MaintenanceRegression Testing

We will look at general techniques, applicable in several phases of testing

Phases (1)

RequirementsSpecify what the software should do

Analysis: eliminate/reduce ambiguities, inconsistencies, and incompleteness

DesignSpecify how the software should work

Split software into modules, write specifications

Checking: check conformance to requirements, using for example conformance testing

Phases (2)

ImplementationSpecify how the modules work–Unit testing: test each module in isolation

IntegrationSpecify how the modules interact–Integration testing: test module interactions–System testing: test the entire system

MaintenanceEvolve software as requirements change–Regression testing: test changes

Testing Effort

Reported to be >50% of development cost [e.g., Beizer 1990]

Microsoft: 75% time spent testing50% testers who spend all time testing

50% developers who spend half time testing

When to Test

The later a bug is found, the higher the costOrders of magnitude increase in later phases

Also the smaller chance of a proper fix

Old saying: test often, test early

New methodology: test-driven development(write tests even before writing code)

Software is Complex

Malleable

Intangible

Abstract

Solves complex problems

Interacts with other software and hardware

Not continuous

Software Still Buggy

Folklore: 1-10 (residual) faults per 1000 nbnc lines of code (after testing)

Consensus: total correctness impossibleto achieve for complex softwareRisk-driven finding/elimination of faults

Focus on specific correctness properties

Approaches to Detecting Bugs

Software testing

Model checking

(Static) program analysis

…

Software Testing

Dynamic approach

Run code for some inputs, check outputs

Checks correctness for some executions

Main questionsTest-suite adequacy (coverage criteria)

Test-input generation

Test oracles

Other Testing Questions

Selection

Minimization

Prioritization

Augmentation

Evaluation

Fault Characterization

…

Testing is not (only) about finding faults!

Current Status

Testing remains the most widely used approach to finding bugsValidation: are we building the right system?

Verification: are we building the system right?

Testing is gaining importance with test-first development and increased reliability needs

A lot of research on testing (part of mine too)This course is not about research

“Schools” of Software Testing

Bret Pettichord described four schoolsAnalytic (a branch of CS/Mathematics)

Factory (a managed process)

Quality (a branch of quality assurance)

Context-Driven (a branch of development)

This course focuses on artifacts, not process

Do you want a guest speaker from industry?

Topics Related to “Finding Bugs”

How to “eliminate bugs” (localize faults)?Debugging

How to “prevent bugs”?Programming language design

Software development processes

How to “show absence of bugs”?Theorem proving

Model checking, program analysis

Testing Topics to Cover

Test coverage and adequacy criteriaGraph, logic, input domains, syntax-based

Test-input generation

Test oracles

Model-based testing

Testing software with structural inputs

Test automation

Testing in your domain of interest?

Summary of the Introduction

Eliminate bugs to save lives and money

“Bugs” may mean faults, errors, failures

Several approaches for detection: software testing, model checking, static analysis…

Software testing is the most widely used approach for validation and verificationWe will cover systematic approaches to testing, based on coverage criteria for various models

Testing is not (only) about revealing faults