Académique Documents
Professionnel Documents
Culture Documents
This premier edition sees tall tales of new code, The main peeps behind the magazine:
stories of agile success, advice on gatekeeping,
real life ghost stories, blog updates, cartoons and Rob Lambert
much more. The Social Tester
http://thesocialtester.posterous.com/
Think of this magazine as your little bit of fun.
Your source of testing stories. Your little bit of Rosie Sherry
downtime. Your little magazine of testing. And The Social Creator
feel free to share this magazine with your friends http://rosiesherry.com/
and colleagues. Ping it around and share the fun.
Joel Montvelisky
There’s no doubt 2010 is going to be an exciting The Anti-E Social Tester
year for software testing. With some great initia- http://qablog.practitest.com/
tives and concepts forming and with technology
moving at such a rapid rate, it’s going to be a Phil Kirkham
challenging yet interesting year for many in our The Social Builder
profession. http://expectedresults.blogspot.com/
For some testers it’s early days, they’re learning A huge thanks and high-fives go to all who con-
the craft and building their skills, getting to grips tributed content and all who worked behind the
with what testing is, finding out more about the scenes proof reading and helping out to make
craft and gaining valuable work experience. For this magazine possible.
Articles
Motivating Staff and Traditional Techniques 6
I’m Not Just A Tester 9
The Ghost 11
The Emperors New Code 13
Do You Have Testing Cred 25
Testing in The Open 32
Testers in the Gatekeeper Role 35
Experience Report 37
Tester’s Diary 40
What is Social Media 44
Blogs
Optimistic Developers, Pessimistic Testers 18
The Added Value of Testers 20
You Want Good, Improvisational Agile training? 22
Distributed Agile 22
Conversations
Twitter Conversation 27
Starting Out As A New Tester 28
Fun
Left Hand Tester 5
Automation 8
Career Planning 12
Do Loop Until 0 19
Bug Trophy 26
A Testing Maze 31
Blame 34
Do Loop Until 0 43
Ads
Classified Small Ads 45
Carnival of Testers
Activity from the blogosphere
The blog-activity-ometer has been very busy in New Faces?
the last 3-4 months. Plenty of output, plenty of
discussion, topic cut’n’thrust and the odd (some- The last few months of 2009 saw a bunch of
times very odd) amusing post. testers starting to blog.
Let’s take a sample from the smorgasbord of Yvette Francino bounced onto the blogging
delights that have been those blog posts. scene in September. She has been writing about
a range of topics and issues connected with her
Testing vs. Checking learning of different test-related areas. In this
post she reviews the progress made since starting
Let’s start at the high-end of the Richter scale... Beyond Certification.
Michael Bolton started writing about the “testing
vs. checking” subject in a number of posts (the The end of September saw Peter step onto the
first one here). This was a topic that generated a blogging scene. He covers a wide range of trig-
lot of discussion, comments (on various sites) and gers and influences in his posts, this look at learn-
blog posts - including approval, disapproval and ing styles and philosophy was a typical example.
I was waiting for the Spanish Inquisition (Monty-
python style of course) but I think Michael has November saw Dave Whalen land on the
escaped the torture of the “comfy chair” so far. blog-o-scene with a look at bug report priority
and severity.
Several people wrote (or tested and explored
their thoughts) on it and aspects relating to it, Seth Eliot started writing in December and has
including Albert Gareev, Martin Jansson, Simon produced a few posts worth checking out. One
Morley, Ben Simo and Trish Khoo. of his latest was a post covering the code cover-
age discussion between Matt Heusser, Alan Page
Fair Play and BJ Rollison. For an enlightening view of the
discussion check his “exciting code coverage”
Articles were written encouraging testers to treat post out.
each other with courtesy, polite-
ness, fairness, respect and to not Rounding off the year of new and
pre-judge. People stepping up recent bloggers John Stevenson put
to encourage this were Lannette finger to keyboard. Another tester’s
Creamer, Lisa Crispin, Catherine critical eye being cast on issues of
Powell, Matt Heusser and Jon the day.
Bach. Worth bearing in mind
when the discussions are getting Random Pick?
heated...
Misko Hevery provided an intro
Weekend Testers to getting started with your own TDD. Get past
those nightmares about TDD!
Probably the most recent new group has been
the weekend testers. Many different people have Anne-Marie Charrett helps you discover if you’re
written enthusiastically about the idea. Check incompetent without knowing about it! Don’t
out James Bach’s post, Bharath’s summary and have nightmares if you know that you don’t know
Parimala Shankaraiah’s experience report. something!
R
ecently I’ve been thinking and writing and a senior statistician to set aside two days to
about the effects of testing standards. brief me and my assistant. There was just us, a
The more I thought, the more convinced flip chart, and gallons of coffee as they talked us
I became that standards, or any rigid processes, through the market they were competing in, the
can damage the morale, and even the problems they faced and their need for better
professionalism, of IT professionals if information from the underwriting and
they are not applied wisely. claims applications with which they
did business.
The problem is that calling them
“standards” implies that they are I realised that it was going to be a pig
mandatory and should be applied of a job to give them what they
in all cases. The word should be needed. It would take several months.
reserved for situations where However, I could give them about a
compliance is essential, eg security, quarter of what they needed in short
good housekeeping or safety critical order. So we knocked up a quick
applications. disposable application in a couple of
weeks that delighted them, and then got to
I once worked for a large insurance company as work on the really tricky stuff.
an IT auditor in Group Audit. I was approached
by Information Services. Would I consider mov- The source systems proved to be riddled with er-
ing to lead a team developing new management rors and poor quality data, so it took longer than
information (MI) applications? It sounded inter- expected. However, we’d got the users on our
esting, so I said yes. side by giving them something quickly, so they
were patient.
On my first day in the new role I asked my new
manager what I had to do. He stunned me when It took so long to get phase 1 of the application
he said. “You tell me. I’ll give you the contact working to acceptable tolerances that I decided
details for your users. Go and see them. They’re to scrap phase 2, which was nearly fully coded,
next in line to get an MI application. See what and rejig the design of the first part so that it
they need, then work out how you’re going to de- could do the full job on its own. That option had
liver it. Speak to other people to see how they’ve been ruled out at the start because there seemed
done it, but it’s up to you”. to be insurmountable performance problems.
The company did have standards and processes,
but they weren’t rigid and they Our experience with testing had shown that we
weren’t very useful in the esoteric could make the application run much
world of insurance MI, so we faster than we’d thought possible, but
were able to pick and choose how that the fine tuning of the code to
we developed applications. produce accurate MI was a nightmare.
It therefore made sense to clone jobs
My users were desperate for a and programs wholesale to extend the
better understanding of their first phase and forget about trying to
portfolio; what was profitable, and what was hack the phase 2 code into shape.
unprofitable. I had no trouble getting a manager
‘Making an observation linked to a decision rule How about all the learning, practicing and im-
that may be performed by a machine…’ proving I want to do?
Have you ever worked with a 3rd party and you How about part of the team?
had to do some testing? Well the 3rd party are
providing you a service aren’t they? In this situa-
tion you could be the customer couldn’t you? By Peter Haworth-Langford
http://007unlicensedtotest.blogspot.com
What about creating some tools to help with your
As designed. www.softwaretestingclub.com
The Ghost
by Luisa Baldaia it was the first time that the
only task they were asked to do
a tool without identifying the
needs.
was Testing. They were a small
team too. Back to the new (reorganisa-
I would like to tell you a story. tion) era, how do we test now?
A story about the time a Soft- I could say we practice a little
ware House starts testing its
“It was so
of Exploratory Testing and a
own software. little of Checking. We work
simple, why
in a month development life
You might think it strange, cycle. A lot of new procedures
start testing? What did they do
didn’t we think
were defined in Quality, like
before? It’s not that strange. In Acceptance Criteria (the point
Portugal many software houses
about it before?
when it is acceptable by the
still don’t do formal testing, at client), Regression Tests and
least the way we think about
And we can
Maintenance tests.
testing.
The whole night before the site went live the The Emperor settled most of his lawsuits with
rogues were awake, playing multiplayer online former employees out of court, got a major book
games against a team in Lesotho, Africa and deal and sold the rights to have it made into a
finally ensuring that all browsing histories and major motion picture.
cookies were removed from their machines and
that anything small enough and not nailed down The rogue developers used the proceeds of their
was in their travel bags. The Emperor himself scam to emigrate to Eastern Europe where they
had organised a massive viral marketing cam- now run a major spam factory.
paign with A-list celebrities endorsing the won-
ders of the new company website. The scripts for And they all lived happily ever after.
the 30 second advertisements were written by the
developers and included the lines, “See how easy
it is to use, how quickly to load”, “You’ll be sur-
prised by our attention to detail” and “Guaran-
teed to work 100% of your time”. Press releases
were issued to all major tv, radio and newspaper
networks extolling the benefits and, ultimately,
return to shareholders. The public launch was
broadcast live from the town square and shown
in 14 different countries in six languages and was
Let’s start on a positive note! Being able to listen Take small steps. Commit to making at least one
carefully in meetings means that when you have contribution to each meeting and the more often
something to say, its going to be relevant and you contribute, the more comfortable you’ll feel.
valuable. It will take time, so be patient with yourself !
To help you feel more comfortable in sharing If it’s possible, choose someone you trust to ask
your thoughts with others, consider this stepped for feedback after the meeting.
approach:
Take whatever feedback they have (and the more
Ensure you read through the agenda and think specific and detailed the better!) and take action!
about the possible things (topics) that might be
discussed. Repeat these steps until you find contributing
easier; taking less conscious effort.
Next, clarify your thoughts on these topics
• Developers are creators, with a natural opti- Optimistic Developer: We’ll fix those bugs later,
mism about making new things and solving when we have time
difficult problems. Pessimistic Tester: We never have enough time to
• Testers are fault finders, with a necessary fix the bugs
skepticism and doubt.
• If developers are the yin, testers are the yang. Optimistic Developer: This build is feature com-
plete
I believe this is a good thing, a sort of checks- Pessimistic Tester: The features exist; some are
and-balances tension that makes for better completely broken
software.
Optimistic Developer: Anything is possible, given
But it does lead to some interesting contrasts... enough time
Pessimistic Tester: Everything has flaws, and
Optimistic Developer: The glass is half full given enough time I can prove it
Pessimistic Tester: The glass is twice as big as
required Optimistic Developer: Of course it will work
Pessimistic Tester: It might work, but probably
Optimistic Developer: This code hasn’t yet been won’t
tested. It’s not known if it has any bugs
Pessimistic Tester: This code hasn’t yet been Optimistic Developer: One last bug fix, and we
tested. It’s not known if it actually works can ship tomorrow
Pessimistic Tester: Fixing this one bug will likely
Optimistic Developer: We are 90% done lead to two more
Pessimistic Tester: We don’t know when we’ll be
done, if ever Optimistic Developer: Stop finding bugs, or we’ll
never be done
Optimistic Developer: We will refactor the code Pessimistic Tester: Stop creating bugs, so I can
to make it better find them all
Pessimistic Tester: They are throwing out the
working code and replacing it with an unknown Optimistic Developer: There’s no need for more
quantity tests
Pessimistic Tester: Let’s just run a few more tests
Optimistic Developer: to be sure
I only changed one line of code
Pessimistic Tester: Optimistic Developer: There is no I in TEAM
The entire system must be retested Pessimistic Tester: We can’t spell BUGS without
U
“It works on
my machine”
Optimistic Developer: Willing to bet that there Optimistic Developer: I’m a Realist
are no more bugs Pessimistic Tester: I’m a Realist
Pessimistic Tester: Willing to take that bet
Source:
Optimistic Developer: Let’s slip these changes in http://www.sqablogs.com/jstrazzere/1819/Opti
now, because I’m starting my vacation tomorrow mistic+Developers,+Pessimistic+Testers.html
Pessimistic Tester: Let’s not
www.doloopuntilzero.com
You Want Good, Improvisational Agile Training? Everyone is “going Agile” these days. Even
though Agile Software Development methodolo-
This ain’t it: gies have been around for awhile, there seems to
be more and more buzz about them, offering a
Laughing. Overcoming embarrassment. Out- plethora of options such as the popular Extreme
and-out goofy behavior. These are not normally Programming (XP) and Scrum methodologies.
the skills managers seek in their developers. But New variations continue to pop up and many or-
some agile development advocates believe these ganizations seem to work in an environment that
skills are critical for successful software projects. is kind of a hybrid of Waterfall and Agile, want-
ing to maintain the discipline of Waterfall and
A growing number of artistically inclined cor- the flexibility of Agile. So what exactly is Agile?
porate trainers are promoting the principles of
the Agile Manifesto with techniques from acting, A team of 17 technologists got together in Febru-
improvisation and other art forms. Such exer- ary, 2001, and authored the Agile Manifesto:
cises attempt to prepare software developers for
changing requirements and other unexpected We are uncovering better ways of developing
occurrences throughout the agile development software by doing it and helping others do it.
process. Developers learn better ways to work Through this work we have come to value:
together and how to put the team before the • Individuals and interactions over processes and
individual. tools
• Working software over comprehensive docu-
If you want to train them how it really works, mentation
you do the same thing I do to train new QA • Customer collaboration over contract negotia-
people: You cut the electricity, pull the fire alarm, tion
and wait by the stairwell with an old American • Responding to change over following a plan
Gladiators padded quarterstaff you bought off
of eBay. That is, while there is value in the items on the
right, we value the items on the left more.
Anyone who makes it out of the building can sur-
vive the ever-changing timelines and predictable, While the various methodologies outline detailed
avoidable, but always sudden “emergencies” that, processes aimed at implementing these guide-
well, emerge. Anyone who knocks me off my feet lines, the key to success lies in effective commu-
or separates me from the quarterstaff gets a field nication, collaboration and teamwork between
promotion to Senior Test Engineer. business folks, developers and testers. Because
communication, collaboration, and teamwork are
Or if you want to simply waste an afternoon, I all more easily accomplished from face-to-face
guess you could hire these Agile training consul- time, typically it’s recommended that agile teams
tants for some entertainment and then go back to are co-located – not just working at the same
work the next day doing it the same way you’ve site, but actually sitting together as they work.
always done it. (I’ve heard tales of XP teams being locked up in
a conference room together until they finished
Source: a project, though the person that told me that
http://qahatesyou.com/word- tends to exaggerate. I’m sure the custodian let
press/2009/02/13/you-want-good-improvisa- them out for biological needs.)
tional-agile-training/
The comprehensibility of your reports and the The information for this article was sourced from
extent and skill of your analysis will also have a the Bug Advocacy 2008 course provided by the
substantial impact on your credibility.” Association of Software Testing. Many thanks to
all the instructors: Cem Kaner, Doug Hoffman,
A lovely example of a tester who has earned Adriano Comi, and Brett Leonard.
good tester credibility is by a Microsoft tester
named Ricky Kurniawan. He describes in his Anne-Marie Charrett is an independent
post how he pursued a multi threaded issue that software test consultant at Testing Times
could only be reproduced on his machine. You and blogs at Maverick Tester.
can read the full story here A Bug Story: Multi-
qualityfrog: @skim A smaller hamster wheel does Follow @skim and @QualityFrog on Twitter.
not make one agile. :D
Automation
for someone else.
This is a very incomplete list
Steve K: I agree. There is but already it’s getting com-
not only one way to get into Simon: Yep, avoid the auto- plicated - so it makes sense for
testing. The approach will vary mation voodoo to start with- If both the new employee and
depending on the environ- the company has some struc- employer to have a way of
ment, people, etc. tured “training” even better keeping track of some of the
- “testing” is such a wide open basics - one of the first review
Simon: I think on-the-job- field that it can be overwhelm- times it could be a task of the
training can sometimes be the ing to start with. new tester to suggest what’s
new tester’s first intro to pair- missing from his checklist (after
testing (actually sitting along- Structured training isn’t 2-3 months?) – so that the
side someone and working thru necessarily courses but maybe guide is personalised..
a problem). Unfortunately for more of a checklist/guide of
some it might also be their last concepts and types of testing Checklist/exploratory: Yes
experience... to experience - this probably agree completely. A very good
helps both the new tester and approach (this works for not-
For me, the biggest benefit for the employer - have some sort new testers also) is to take a
the new tester is getting their of record of types that have scripted test (not necessarily
hands on a real problem - been covered and what hasn’t automated) and ask the tester
actually practising. They can been covered. Some project if this test is still relevant, does
make mistakes as well as find iterations might need more it miss anything, where/how
bugs in the app. Some bugs of one type than another - it could it be improved - and if it
might be due to mishandling/ might be a GUI change, or is sufficient, why is it sufficient?
configuration due to them input field modification, some
being new - but this is all good network protocol introduction, Here the mentor needs to be
learning sources. etc, etc helping/pushing the new tester
to think about what’s going on
So the checklist might start on and why...
a high level Static Analysis/
Is it an issue? www.softwaretestingclub.com
I can’t think of an environ- points. ;-)) - but let’s just dive in and
ment where not jumping in see what happens. I’ll pick up
with some OJT is not a good Simon: Yes, faults and mis- some tips by observing how
thing - obviously I’m not in- takes (whether in the app or others do it - and see if they
cluding any life-critical systems made by the tester) are es- can be applied to me (or used
here - although it could still sential for learning. I know I by me). “Waving” seems to be
work there. The new tester will wouldn’t have learnt so much a low cost or easy “entry-level”
just have their work checked about the systems I have app so there’s no need to read
by their supervisor/mentor/ worked on at the start if they a manual or whatever - espe-
helper more... were fault-free. cially since the look’n’feel is
similar to parts of other apps
OJT?
Interesting about alpha/beta and there are parts that are in-
testers - like you say some tuitive - but that doesn’t mean
want to be first in line or (early I’m doing it “right” or that I’m
Steve K: What is OJT? adopters as I think marketing making the most of it....
peeps classify them) but they
Simon: OJT: On-the-Job- don’t always want to be find- Now back to testing - there’s a
Training ing the problems. Aren’t some bit of this that’s an analogy to
more “fair-weather alpha/ learning about testing...
Simon: So, what are the pre- beta testers” (only want it if it
requisites, what’s needed for a works - it’s a kudos thing)? But, Simon: So for a new tester,
new tester to get going? Curi- I think in general when you’re what are good resources on-
osity, willingness to learn, not employed to be testing then line? Are some starting points
being afraid of making mis- you have to want it, otherwise better than others? Can some
takes??? Are these good start- it’s not going to work. places be more intimidating
ing points before later making than others?
more “academic” demands Simon: Think about courses
about critical thinking/reason- and learning new topics - some Simon: On-line resources /
ing? people want to do a course/ learning aid ? Are there some
tutorial before practising, some that are better than others?
Steve K: I think by default want to get a feel of a new
everyone has a level of curi- topic “hands-on” before taking What about twitter? This
osity and eagerness to break a course and some don’t want seems to have a life of its own.
software. It’s quite interest- the course but want some “on- I wouldn’t recommend it for a
ing to see how many people the-job” training - then there new tester (or any tester re-
want to sign up for alpha/beta are others that just want to try ally) as a resource for learning
software to get the chance to it for themselves, let it sink in about articles.
try out new products/services and see what they make of it -
and perhaps a bonus to be able “so called exploring...”
to brag about it. I think “new
tester” testers should be pas- The thing is that for some top-
sionate, willing to learn, and ics/areas one type will work
definitely not be afraid of mak- for someone and then for a
ing mistakes (as you’ve men- new topic maybe a different
tioned). While some people method is better for that same Rob: For me Twitter has
think failing is a bad thing, it’s person... opened up a whole new hive of
part of the growing pains to contacts and feeds where I get
become a better tester. We’re Take “Google wave” - I’ve loads of new articles/ideas and
human after all. no idea if I’m doing this in a concepts. I’m not sure I would
good way or not (I’m gonna have come across these had it
I think those are good starting stop/pause soon for fear of not been for Twitter.
this turning into a monologue
Luck Factor
from a new tester’s perspective works in other forums also -
- it could be overwhelming / generally testers of all levels
information overload. are approachable.
But, as with most things I’d Simon: Something that just
recommend a “suck it and see” Rob: This is such a good occurred to me - there’s a
approach - find out yourself benefit of Twitter. The ability “luck” factor for the new tester
what you’re comfortable with. to interact and chat (albeit it also: luck that they begin in a
in small sentences) with people “good” work environment that
Rob: Yeah. New testers could you would never normally supports and encourages them.
become swamped by the mass have been in touch with is fab] Can be intimidating to come
of noise. Takes a while to filter in and think that you should
Blogs
it out before it starts to make be “questioning” and trying to
sense. make sense of everything that
the unit/org is producing/de-
Steve K: I agree with the both Simon: Blogs? It’s hard to veloping - especially if there’s
of you. Twitter has opened up recommend this as a way of some developers that aren’t too
an opportunity to listen and learning about testing - many happy with questioning from a
talk to other passionate testers/ times they can be “parts of new tester... Thoughts?]
developers and insurmount- conversations/topics”
able amount of information in Rob: Couldn’t agree more. In
articles, blogs, presentations, Rob: Agreed. Blogs are often fact some testers can get stuck
etc. The signal-to-noise ratio is very personal, often incredibly in a “bad” environment and
definitely overwhelmingly for a misleading and sometimes very not know that “better” ones
new tester. Once the new tester political. Saying that though, exist. It pays to move roles
familiarizes test concepts, he/ they are invaluable sources of every so often. Keeps your
she can start diving into the answers to questions skills sharp and builds on your
testing world in Twitter. experience. After all, no two
Simon: Q&A and discus- companies, methodologies or
But where it is powerful (but sion forums are a good way products are the same.
still a bit random) is that you in - generally questions can
can contact someone who be fired off and usually get
wrote an article or is knowl-
Email rob@softwaretestingclub.
Provide an IRC channel to talk about your You have to add as much information as possible.
project. If your testing community is big enough, Your testers should be focused on testing and
open a channel just for testing activities. Go there only testing during the day to make the most of
often. Answer questions. Don’t make your testers it.
feel lonely in their task.
Here’s some ideas for what to include on the
Meetings are also a good way to keep up com- Testing Day website:
munication between developers and testers.
Collaborators know when they are going to meet • Goal of the Testing Day. Are we testing any
and there is an agenda (which should be open new features? Any specific hardware? Include
to add items to discuss), which helps driving the any relevant information and point to docu-
conversation. mentation about the feature plus relevant
installation files etc
“A nice way to encourage test- • Include information about when this testing
day is happening. Is it over a weekend? Only
ing during an important time of a day? Which time zone?
the cycle are organized Testing • What is the main medium of communica-
tion for that Testing Day? Sometimes Testing
Days.” Days are the first contact testers have with
your project, so try not to assume previous
knowledge.
• Announce it properly to ensure a maximum
The common place to have these regular meet-
success. Blog about it, tweet about it, an-
ings is the IRC channel of the project (of the
nounce it in testing and open source forums.
specific channel for testing activities, if you have
More is always better.
created one). If you have testers around the
• During the Testing Day ensure you are
world, swapping between time zones to schedule
available (at least during your core working
the meeting is always a good idea. It would allow
hours). Be at IRC, answer peoples questions,
every member to attend at least one meeting.
keep tweeting about it and promoting it.
Fixed. www.softwaretestingclub.com
Experience Report:
Introducing Exploratory Testing
By Simon Morley
So you’ve heard of Explorato- should do something, how we at interactions between NE’s as
ry Testing (ET). You’ve heard should do it and the when was well as the core use-cases in the
the PR and thought it sounds ... “as soon as possible, if not NE system under test (SUT).
interesting. Yeah, that sounds sooner”. The “if ” we should The main reason for introduc-
like something I’d like to try. do something was in the con- ing the activity was that the
text of the other testing that target feature interactions were
Maybe you’ve read some of had already been conducted due to be tested in the “fewer
the literature in blogs (lots of on the feature. simulated interfaces” environ-
people talking about it), web- ment.
sites (for example the context- Don’t go overboard on After some sessions involving
driven school?), books (James different types of questions,
Whittaker came out with one the “process” or struc- investigations and follow-up
in the autumn) or some other ture of the work. Start sessions we came to the con-
buzzword being discussed in gently, learn and adapt clusion that there were some
conferences or on twitter. “interesting” cases/scenarios
the “process” to your to look at. These were areas of
Whichever way that you’ve own needs. the functionality and system
come across ET how do you configuration where we saw
get started with it and intro- Time for a new intro? that we could add some value -
duce it into a team or organi- Analysis... we identified some “risk areas”
sation? Well, this is something - areas where we thought “well
I did recently - so think of this We set up a small team that we don’t know if this will work
as a case-study or experience started analysis on the target when we plug in this piece of
report. feature - what it meant in hardware”.
an isolated node and what it
Why introduce it now? meant in a larger system. This The whole analysis phase was
was an opportunity for us to looking at different aspects of
There was a need to test/assess get to know the feature and the problem - understanding
a feature in a more complex other aspects of the environ- the end-user requirements,
environment (middle picture ment (there were multiple simplifications we were making
below) as there were problems features changing at the same in the environment (e.g. aspects
doing it in the full customer- time.) Besides the test team we to consider by simulating parts
like environment (right-hand involved various designers and of the network) and what it
picture below.) Here was an system managers to under- might cost to test in the envi-
opportunity to re-use some stand different aspects of the ronment we wanted.
previous experience of ET and “problem”.
introduce a “structured active The analysis ended with a “go
learning/test” activity with the With regard to the picture decision” - we had an OK cost
emphasis on quick start-up. below we were looking at for the simulation effort and
the middle environment - a we’d justified the need for the
Our goal was to assess the “small” network where a testing by identifying some
big picture (below, as well as single network element (NE) “risky” areas that we thought
smaller individual configu- was implementing a range of we give us some benefit to
rations) and conclude if we features and we were looking explore.
FAIL. www.softwaretestingclub.com
work processes so that they are It’s infectious! vironment/configuration and
more ET-like. Take as big or how that relates to the main
little steps as you feel comfort- Once you start a team look- use-cases
able with. ing a problem from a new
angle it doesn’t take long for Don’t put more effort in the
Set out your own objectives the ideas to be self-generating. up-front test design than is
and following-up “how” you Our testers have a wide range needed to get started
meet them (you may decide of freedom in following up
that something is not working “hunches” but this really took Experience in ad-hoc/ET is
and it’s time to re-think how to off in this exercise. The team good, but getting started and
look at the “problem”.) would change tack during the learning from mistakes is just
test session quite dynamically. as good
Be sensitive with the team We evolved a set of target tests
dynamics. Relate what you’re that we wanted to perform but Document what you do - the
doing to new ground - this has these were added to and modi- whole activity is a feedback
worked for me in the past. fied quite dynamically. loop - you may discover some
use-cases are no longer so
In the beginning I didn’t have
all the details about how I “Not everyone was important for the scope you’re
working in
wanted to run the ET trial. comfortable with a
That’s OK - it’s more about
being able to adapt as you go. less-structured ap- As Nike say, “just do it”.
December 7th
It turns out that we are sup-
posed to create fridges for the
storage of ice cream in super-
markets and they have hori-