Agile Testing Talk Nov2006

Agile QA/Testing
Elisabeth Hendrickson
Quality Tree Software, Inc.
www.qualitytree.com
Last updated November 20, 2006
Contents
• Agile overview
• How traditional testing practices evolved
• Agile and traditional projects contrasted
• Agile test practices evolving
Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 1

Manifesto for Agile Software Development
We are uncovering better ways of developing software by doing it and

helping others do it. Through this work we have come to value:
Individuals and interactions over processes and tools
Working software over comprehensive documentation
Customer collaboration over contract negotiation
Responding to change over following a plan
That is, while there is value in the items on the right,
we value the items on the left more.
Kent Beck, Mike Beedle, Arie van Bennekum, Alistair Cockburn,
Ward Cunningham, Martin Fowler, James Grenning, Jim Highsmith,
Andrew Hunt, Ron Jeffries, Jon Kern, Brian Marick, Robert C. Martin,
Steve Mellor, Ken Schwaber, Jeff Sutherland, Dave Thomas
© 2001, the above authors
This declaration may be freely copied in any form,
but only in its entirety through this notice.
History of the Agile Manifesto

The signatories of the Agile Manifesto gathered at The Lodge at Snowbird ski resort
in February 2001. Among the group were representatives from Extreme
Programming, SCRUM, DSDM, Adaptive Software Development, Crystal, Feature-
Driven Development, and Pragmatic Programming.
The result was the Agile Manifesto, a set of statements that synthesize the core
values of the various Agile methods.
Agile Began Before 2001
Some people think that the Agile Manifesto kicked off the whole Agile Movement.
This view is supported by the fact that from 2001 to today, there have been dozens
of books published on Agile development under titles relating to XP, Scrum, Lean,
Agile, and Adaptive.
However, although the creation of the Agile Manifesto may have been a turning
point, it wasn’t the beginning. The authors of the manifesto had been working in an
Agile way for years before they articulated their shared vision for what Agile meant.
Even before that, in the 1980’s Tom Gilb and Barry Boehm published works on
Spiral and Evolutionary software development.
And in the 60’s and 70’s, software pioneers were Agile out of necessity. Gerald
Weinberg notes that they used test-first development on the Mercury project,
designing and implementing tests before code. For more information, see
http://c2.com/cgi/wiki/wiki?HistoryOfIterative.

Examples of Agile Methods
Lean Scrum
Lean manufacturing Lightweight
concepts applied to management
software framework.
development.
Crystal Extreme
Lightweight set of Programming (XP)
development Rigorous set of
practices. practices designed to
keep both the code
and team agile.
Agility and Agile Methodologies

“Agile” is a great buzzword. Who doesn’t want to be agile? No one says, “thanks,
I’d rather be inflexible and slow to respond.” The problem with buzzwords is that
people invoke them without understanding what the terms really mean. “Agile” is a
relentless focus on providing a continuous stream of business value, even in the face
of constant change.
For more details, see:
• The Agile Manifesto: http://www.agilemanifesto.org
• The Agile Alliance: http://www.agilealliance.org
• The Declaration of Interdependence: http://www.pmdoi.org/
• The Agile Project Leadership Network: http://www.apln.org
Agility is usually achieved by adopting one or more Agile methodologies.
Mixing and Matching Methodologies
Not all Agile methodologies address the same concerns. For example, Lean and
Scrum are project-centric in that they provide managers with tools to monitor,
prioritize, and control projects without specifying anything about how the software
is built. By contrast, Crystal and XP are developer-centric in that they provide a set
of good practices for a development team and less for a project manager.
Some organizations have tried combining methodologies with great success. One
common combination is XP + Scrum.

Lean
Principles: Tools:
• Eliminate waste • Seeing Waste
• Amplify learning • Value Stream Mapping
• Decide as late as possible • Feedback
• Iterations
• Deliver as fast as possible
• Options Thinking
• Empower the team
• Last Responsible
• Build integrity in Moment
• See the whole • Queuing Theory
• Cost of Delay
Source: Mary Poppendieck, • Testing
Tom Poppendieck. Lean
• Refactoring
Software Development: an
Agile Toolkit. Etc. (22 tools in all)

Scrum
Scrum Scrum
Team Master
Daily
Scrum
Product Sprint
Owner Backlog
Sprint
30 Days
Product
Increment
Product
Backlog
Adapted from the original at
http://www.controlchaos.com/about/
The Daily Scrum

Three questions each team member answers:
1. What have you done since the last Scrum?
2. What do you plan to do before the next Scrum?
3. What is getting in the way?
The Scrum Master Removes Obstacles
At one morning standup meeting, Elisabeth learned that one of the team members
would not be able to deliver the code as promised. “What’s getting in the way?”
Elisabeth asked. The team member reported that an executive had asked her to help
resolve a tech support issue.
Resolving the tech support issue was important. But the executive who pulled the
team member did so without taking into account the cost to the project. Elisabeth
worked with the executive to find another resource who could help tech support,
freeing the team member up to meet her commitments.

Crystal
“ The lead designer and two to seven other developers

in a large room or adjacent rooms,
using information radiators such as whiteboards and
flipcharts,
having easy access to expert users,
distractions kept away,
deliver running, tested, usable code to the users
every month or two (quarterly at worst),
reflecting and adjusting their working conventions
periodically.
Source: Alistair Cockburn, Crystal Clear. Available from

”
http://alistair.cockburn.us/crystal/books/alistairsbooks.html

Extreme Programming (XP)
Values: Practices:
• Communication • The Planning Game
• Simplicity • Small releases
• Feedback • Metaphor
• Courage • Simple design
• Testing
• Refactoring
• Pair programming
• Collective ownership
• Continuous integration
• 40 hour week
• On-site customer
• Coding standards
Source: Kent Beck, Extreme Programming Explained.

Agile Synthesized
Adapt Collaborate
Embrace change. Give up on Co-locate team members.
“managing” or “controlling” Whole team thinking. Get
it. close to the customer.
Get Feedback Communicate
Test early, test often. Information radiators. Big
Everyone tests. Customer visible charts. Daily
“accepts.” standup meetings.
Deliver Value… …Continuously
Do important features Short timeboxes
first. Reduce waste. (“iterations” or “sprints”)
Reduce overhead. Reduce
rework.
Agile Expectations
Change is inevitable.
The whole team, including the programmers, testers, customers, is responsible for
the outcome.
Everyone on the team is accessible and actively communicating throughout the
project.
Programmers test early, often, and aggressively.
The whole team actively solicits feedback.

Calling It “Agile” Doesn’t Make It So
This is NOT Agile:

1. Compress the
schedule
2. Toss out the
documentation
3. Code up to the last
minute
The organization may gain
short term speed but at the
cost of long term pain.
Agile Lip Service

Unfortunately, as often happens with great buzzwords, some teams claim they’re
“going Agile” when all they’re really doing is compressing the schedule, tossing out
the documentation, and coding up to the last minute. As Abraham Lincoln said, “If
you call a tail a leg, how many legs does a dog have? Four. Because calling it a leg
doesn’t make it a leg.” Calling it Agile doesn’t make it Agile.

How Traditional Testing Practices Evolved
With great optimism and the best of

intentions, The Project Plan is announced:
Analyze Design Code Test/Bug Fix
Requirements Completed Code

handed off to Dev handed off to Test Release
Phased Development in an Ideal World

If we lived in a perfect world, it would be possible to know all the requirements up
front. We’d get them right every time. We’d somehow uncover all the implicit
assumptions and ambiguities. We’d have the power of perfect prediction to know
every permutation of configurations and sequences that will occur in production.
In a perfect world, our designs would be perfect the first time. We’d know exactly
where the performance bottlenecks were going to be, and would create a perfect
architecture to compensate accordingly.
In a perfect world, our code would always do what we wanted it to, the first time.
We’d have perfect understanding of all the 3rd party libraries we rely on and would
use them correctly, and make the calls in the right sequence.
In a perfect world, whatever small issues remained at the end of the cycle would be
easy to find and fix in plenty of time before the release. The fixes would be minor,
requiring no major refactoring, no huge architectural changes.
Alas, we don’t live in a perfect world. Technology, being made by humans, is
especially imperfect Inevitably, unexpected discoveries lead to rework and
schedule slips. Testers have learned that no matter what the project plan says, the
reality is likely to be…

How Traditional Testing Practices Evolved
Inevitably, The Project Plan is revised:
Analyze, Design, & Code Test/Bug Fix
Completed Code
handed off to Test
Release
The Problem of Late Testing

When the bulk of the test effort is squeezed into the last few months or weeks of a
long release cycle, the time for testing evaporates with each successive schedule slip
in preceding phases. The testers know that the less time there is to test, the less they
will be able to test. And the less testing they can accomplish, the less the
organization will know about the software before it’s released. Less information
means more risk.
For years, testing organizations have sought ways to do more testing in less time.
Common approaches include: test automation, outsourcing, insourcing, bug bashes,
and internal or customer beta tests to augment the internal test effort. Let’s be
honest: while some organizations have seen some gains from these practices, most
organizations are still overwhelmed with too much to test and too little time in
which to do the testing. Even if the testers are able to complete the testing, there
often is no time left to fix the bugs testing reveals.
Although these problems are widely recognized, real solutions have proven elusive.
Yet we’ve developed a number of practices aimed at controlling the end-of-release
chaos.

The Result: Practices Intended to Control Chaos
Traditional testing practices attempt to manage

the chaos (or at least avoid the blame):
– “Last Defender of Quality” stance
– Strict change management
– Detailed preparation and up front planning
– Heavyweight documentation suitable for outsourcing
the test effort
– Strict entrance and exit criteria with signoffs
– Heavyweight test automation focused on regression
– Attempts at process enforcement
These practices are anti-Agile. Can independent
test groups adapt in an Agile environment?
The Solution to the Late-Testing-Problem

It’s a deceptively simple solution. If delaying testing until the end of the release is
risky…don’t delay the testing until the end of the release. Easy to say, hard to do—
unless you’re in an Agile environment that supports early testing. Even then, it can
be difficult to get testers involved from the beginning.

Agile = Continuous Stream of Value
Iterative approaches mean we can trade features for

time instead of sacrificing quality.
Iterations
Completed, Releasable Features:

Designed, Coded, Tested
Necessary Conditions for Adopting Agile

Here’s how to tell if an Agile approach would work in your environment:
• The organization is willing to embrace agility as defined by the Agile
Manifesto. Saying “Be More Agile” or “Go Faster” isn’t enough. The team
must be more interested in usable feedback than comprehensive
documentation or detailed plans, and more interested in collaboration than
signoffs.
• Everyone tests, not just designated testers. Agile teams are “test infected.”
(An important side effect of teams being test infected: the software is built to
be testable from the beginning.)
• The whole team is responsible for quality, not just the testers or people with
“Quality” in their title. Which are you more likely to hear: “How did you miss
that bug?!?” or “What can we do next time to prevent something like that
happening again, or detect it earlier?” When everyone is responsible for the
outcome, there’s more collaboration and less finger-pointing.
• Managers focus on fixing problems, not blame. Agile practices don’t provide
CYA paper trails and are unlikely to succeed in a high-blame, high-fear
environment.

QA Becoming Agile: Shifting Roles
“Fear not!
I’ll protect
you!”
“Hey, would
this help?
from last line of
defense… …to team support
QA Cannot Be Agile in Isolation

The practices described in this talk don’t work when implemented in isolation. As
we’ve seen, traditional test practices evolved to support traditional phased
development processes. If you’re in a traditional environment, stick with the
traditional test practices: they’re a better fit for your context.

Becoming Agile: Delivering Value
Increase the rate

of delivery
(usually with
smaller increments)
Increasing the Rate of Delivery

Life in an un-agile company:
“How long until we can deploy the new Mobile client?” asked Jamie, the Product
Manager. “I heard from Jill that the coding is done.”
“Yes, the Mobile features are coded,” responded Gerry, the Development Manager.
“But we can’t deploy for another six months.”
“Six months?!?” Jamie’s jaw dropped. “Why?!”
“Because the rest of the release won’t be ready for another six months.” Gerry
replied. “That’s the soonest anything can go out.”
Six months is an eternity in today’s marketplace. In six months, the competition
may have leapfrogged the current solution and snagged the majority of the market
share. Speed is of the essence.
One way some organizations expedite the release of critical fixes or features is with
a patch release. But patches come with a high cost. Usually people have to be
pulled off their regular work to do a patch, test systems must be reset, and work on
other releases comes to a screeching halt. Once the patches are done for a given
release, the changes must be merged into one or more additional code lines. Code
base management becomes a nightmare.
Agile teams use smaller increments and just-in-time collaborative planning to
ensure critical changes can be released sooner. If something is important, it can be
moved into an earlier release or iteration. It may displace another feature that isn’t
as important, but it won’t disrupt an entire release. The result: faster delivery.

Becoming Agile: Increasing Feedback
Increase the rate

and quality
of feedback
Shortening Feedback Loops

How long does the team have to wait for information about how the software is
behaving? Measure the time between when a programmer writes a line of code and
when someone or something executes that code and provides information about how
it behaves. That’s a feedback loop.
If the software isn’t ready to test until the very end of a long release, the feedback
loops will be extended and can be measured in months. That’s too long.
Shorter feedback loops increase Agility. Fortunately, on Agile projects the software
is ready to test almost from the beginning. And Agile teams typically employ
several levels of testing to uncover different types of information.
• Automated unit tests check the behavior of individual functions/methods and
object interactions. They’re run often, and provide feedback in minutes.
• Automated acceptance tests check the behavior of the system end-to-end.
They’re typically run on checked in code on an ongoing basis, providing
feedback in an hour or so.
• Manual regression tests take longer to execute and, because a human must be
available, may not begin immediately. Feedback time increases to days or
weeks.
It’s no wonder Agile projects favor automated tests. Though manual testing,
particularly manual exploratory testing, is still important, even on Agile projects.

Becoming Agile: Reducing Waste
Reduce waste
Waste in Software Development

In Lean terms, waste is anything that does not add value for the customer. Waste is
not always something physical that is thrown away; inefficiencies are a source of
waste as well. For example, “motion” is considered waste in manufacturing terms.
Software waste can include:
• Half-done features that are ripped out prior to release
• Unneeded features
• Excessive or unmaintained internal documentation
• Rework such as bug fixes/patches
• Invalidated test results such as test results for an invalid configuration or data
set
• Time spent waiting for information, deliverables, decisions, or anything else
that blocks progress
See Lean Software Development by Mary Poppendieck and Tom Poppendieck for
more information on reducing waste on software projects.

Traditional & Agile Projects Contrasted
Traditional Agile
Change Manage & control it Change is inevitable
Planning Comprehensive up Plan as you go
front design
Document- Verbose Only as much as
ation necessary
Handoffs Formal entrance and It’s not a relay race:
exit criteria collaborate
Test System-level, built by All levels, built by

Automation tool specialists, anyone, an integral
created after the code part of the project
is “done”

Embrace Change: Plan for Maintainability
When creating test artifacts…

Change
• Minimize duplication.
Planning
Thought exercise: if a feature were
removed from your software, how
Documentation
many artifacts would have to be

updated?
• Use tools designed for change.
Hint: if the vendor says “stabilize
Handoffs
the interface first,” run away!

Automation
A Traditional Tester on an XP Project

A tester recently commented to me, “I’m working with an XP team and they’re
driving me nuts! How am I supposed to design, document, and execute a full set of
tests for the product in every 2 week iteration?!? They don’t give me the
specification until I nag them about it! And then in the next iteration, everything
changes!”
This is a tester struggling with embracing change. He wants everything laid out
neatly before he begins his analysis. He wants to do his test design in one fell
swoop. He’s having trouble adjusting his practices for his new context.

Planning: Favor Informal, Collaborative Tools
Informal Whiteboards
Change
Sticky Notes
Planning
Index Cards
Documentation
Wikis
Checklists
Formal Databases
Handoffs
Gantt/PERT Charts
Comprehensive, Polished
Automation
Documents Created from a

Standard Template
A little planning is good. More is not better.

Plan for the current iteration. Design tests for the features or stories to be done in
the near term. Speculative planning means rework.
Use catalogs of reusable tests. Don’t redesign the same tests over and over.
Instead, keep reusable checklists.
Keep an up-to-date, prioritized risks list. What kind of information are the testers
looking for? The risks list covers it.

A Lightweight Example: Wikis
Change
Planning
Documentation
Handoffs
Automation
Built in Twiki with the Table plugin installed. See

http://www.twiki.org
Test Documentation and Test Tracking All in One

This Wiki page provides an example of leveraging one document for multiple uses.
Here we see both the test design and the test results tracking all in one place.
Reporting test status on this project took about 15 minutes total per week.
Tips for Lightweight Documentation
Capture the essence, not the details. Step-by-step instructions cost time without
providing value (usually).
Point to other project documents. If it’s in the user guide, requirements, specs, etc.,
leave it there.
Centralize generic tests in a checklist. Try this: count the number of times
common condition, like invalid dates or null strings, are documented in the test
docs. More than once is too many.

Reduce Handoffs: Integrate Test Efforts
Testing is not a phase. It’s a way of life.

Change • Agile teams are test infected.
Planning The question, “How will we test it?” is
as important as “How will we build it?”
Documentation
• Co-locate testers and programmers.

But sitting side by side does not
ensure communication.
• Track testing status and
Handoffs
programming status all together.

Show tests run-passed-failed together
Automation
with features/stories done and left to

do.

Test Automation on Agile Projects
• Use different types of automated tests

Change for different purposes:
Planning – xUnit tests for unit testing
– FIT/Fitnesse or Domain Specific
Documentation
Languages for acceptance testing.

• Collaborate with programmers on test
infrastructure code.
• Use test automation to support early
Handoffs
exploratory testing.
Automation
The Difference Between Unit Tests and Acceptance Tests

Teams that attempt to substitute unit tests for acceptance tests, or vice versa, often
run into trouble. These are two different types of tests, seeking answers to two
different types of questions. Unit tests answer the question, “does the code do what
the programmer intended?” while acceptance tests answer the question, “does the
system do what the Customer wants?” Unit tests detect unintended change at a
granular level (function, object, method). Automated system or acceptance tests
should cover end-to-end sequences. Don’t substitute one for the other.
Leveraging Test Code for Exploratory Testing
Traditional test wisdom says we can’t start testing a feature until it’s accessible
from an external interface (like a GUI). But we don’t have to wait. Test automation
can facilitate manual exploration.

Agile Testing Practices Evolving
Within an Iteration
Automated Manual
Automated
Acceptance or Exploratory
Unit Tests
Story Tests Testing
Define “done”, Drive design,

Provides
represent represent
additional
executable executable
feedback
requirements specifications
Unit Testing
• Done by developers, usually with an xUnit framework, often as a result of
practicing Test Driven Development (TDD)
• Supports the development process
• Unit test suites represent executable specifications
Automated Acceptance Testing
• The result of a collaboration between the developers and business stakeholders
• Often implemented in a FIT-like framework or Domain Specific Language
• Acceptance test suites represent executable requirements
Exploratory Testing
• Provides additional feedback and covers gaps in automation
• Necessary to augment the automated tests

Further Reading…
Beck, K. (1999). Extreme Programming Explained:

Embrace Change. Addison-Wesley.
Cockburn, A. (2004). Crystal Clear: A Human-
Powered Methodology for Small Teams.
Addison-Wesley.
Crispin, L., & House, T. (2002). Testing Extreme
Programming. Addison-Wesley.
Poppendieck, M. & Poppendieck, T. (2003). Lean
Software Development. Addison-Wesley.
Schwaber, K. & Beedle, M. (2001). Agile Software
Development with SCRUM. Prentice Hall.

Acknowledgements
Many thanks to early reviewers of the

ideas presented here: Brian Marick,
William Wake, Jonathan Kohl, Jeffrey
Fredrick, Daniel Knierim, Marc Kellogg,
Danny Faught, Ron Jeffries, Hubert Smits,
Rob Mee, Sherry Erskine, Amy Jo Esser,
Gunjan Doshi, Dave Liebreich, Janet
Gregory, Chris McMahon

Agile Testing Talk Nov2006

Uploaded by

Copyright:

Available Formats

Agile Testing Talk Nov2006

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Agile Testing Talk Nov2006

Uploaded by

Copyright:

Available Formats

Agile QA/Testing

Last updated November 20, 2006

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 1

We are uncovering better ways of developing software by doing it and

History of the Agile Manifesto

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 2

Agility and Agile Methodologies

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 3

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 4

The Daily Scrum

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 5

“ The lead designer and two to seven other developers

Source: Alistair Cockburn, Crystal Clear. Available from

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 6

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 7

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 8

This is NOT Agile:

Agile Lip Service

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 9

With great optimism and the best of

Analyze Design Code Test/Bug Fix

Requirements Completed Code

Phased Development in an Ideal World

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 10

Inevitably, The Project Plan is revised:

Analyze, Design, & Code Test/Bug Fix

The Problem of Late Testing

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 11

Traditional testing practices attempt to manage

The Solution to the Late-Testing-Problem

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 12

Iterative approaches mean we can trade features for

Completed, Releasable Features:

Necessary Conditions for Adopting Agile

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 13

QA Cannot Be Agile in Isolation

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 14

Increase the rate

Increasing the Rate of Delivery

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 15

Increase the rate

Shortening Feedback Loops

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 16

Waste in Software Development

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 17

Test System-level, built by All levels, built by

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 18

When creating test artifacts…

many artifacts would have to be

the interface first,” run away!

A Traditional Tester on an XP Project

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 19

Documents Created from a

A little planning is good. More is not better.

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 20

Built in Twiki with the Table plugin installed. See

Test Documentation and Test Tracking All in One

Copyright (c) 2004 - 2006, Quality Tree Software, Inc. 21

Testing is not a phase. It’s a way of life.

• Co-locate testers and programmers.