thoughts from the test eye

This years EuroSTAR took place in den Haag, a city that had quite some rain, but also beautiful autumn leaves, and big churches.
The theme of the conference was “The Future of Software Testing”, and a recurring image was a traffic sign informing that the future is next turn to the right. I always thought the future was right ahead of us.

James Whittaker started with a really good and entertaining keynote where he argued that software is extremely important, and that the standard software testing will disappear (see http://blogs.msdn.com/james_whittaker/ for details.) On the other hand, he didn’t mention users or value (as Michael Bolton pointed out at the next sessions (sic!) questions), and he didn’t mention that testers do more than just verify requirements on different environments. We perform low-level and high-level validation and verification simultaneously.

Next I saw Derk-Jan de Grood who presented a mix of Exploratory Testing and structured process-focused testing. It didn’t feel revolutionary, since we Swedes have known about both these concepts for quite some time now.
The Austrian Strasser was a pleasant surprise. He told about Borland’s route from V-Model to Agile, and he stressed the people aspect many times. E.g. the teams themselves could decide if they would be split in smaller groups, and how. They have come a long way, and has a lot of unit test collaboration between developers and testers. “1 + 1 = 3”.
Tuesday ended with a keynote from the Google guys behind Testing On the Toilet (http://googletesting.blogspot.com/2007/01/introducing-testing-on-toilet.html) who has a really good story to tell. And it is a lot bigger than the Toilet, they have done a lot of things for more than three years in order to raise the awareness about testing at their company. A big accomplishment that I think is difficult to repeat at companies that don’t have “20% Time” (At twenty percent of your working time on Google you can do something completely outside your normal activities.)
After the presentation they were asked: “Do you have any objective metrics that can show your results?”
And I was really happy about the answer: “Metrics are difficult, there are so many factors involved. We rely on anecdotal information.”
They also stated: “There will never be a time when you don’t need a solid QA group.”
Wednesday started with Randall Rice sharing his thoughts on the trends that may shape the near future. Some general and good information, but not so much to write home about.
He recommended howtosplitanatom.com for articles on Web 2.0/3.0 and he said that a drawback of SaaS (Software as a Service) is that customers won’t have any control over which version they are running.
Julian Bensaid talked about outsourcing, and how to avoid, and quickly get out of “The Damage Zone”, the phase when thing are going a bit too bad for too long. He stressed “One Team” and said that “You have to innovate to make outsourcing successful.” “Devil’s Square”: quality, time, functionality, money. At question time, he let the audience answer their questions!

Then it was time for Michael Bolton’s double session workshop on Heuristics, which was the next best event for me at the whole conference. He really knows how to talk, he has important things to say, and the exercises stimulated good conversations with the other attendants. (The results are available at http://www.flickr.com/photos/23076491@N02/sets/72157608997920373)
Also, a good presentation stimulates your thinking to new ideas; and I thought of a new definition: “A good tester knows when to break the rules” and I’m wondering if some very good heuristics could be created at your company, where you have the details.
“Maybe testing isn’t supposed to be easy”, “We have to look complexity in the eye”, and he recommended positivedeviance.org.

Next it was my own presentation: Testing is an Island – A Software Testing Dystopia(link!). It went really well, except for one moment when I just couldn’t find any words. I excused myself, took a sip of water, and just started talking again; and after that I really can’t remember exactly what I said. I remember I said the important things, that I referenced some earlier presentations in the way I intended, but I can’t remember any details. I’m pretty sure I explained something in a completely new way that was really good, but I don’t know what it was. I see this as a good thing; my mind was in a state of flow.
Of course you only hear from the positive ones, but it seemed like people agreed that the human element is very important for us testers. I was also happy to hear that my presentation was very different from the other sessions, and that they like this different, thought-provoking angle.
I guess the introduction movie Mårten Ivert and Henrik Emilsson did was very helpful, just because it is so good!
One question afterwards was if/how I could convince people who think that emotions is a bad thing when testing. I said something about not having any proof, but that I might try by saying that we like emotions in life; so why shouldn’t we have them at work as well (Work is a part of our life.) Testers should also in some sense mimic users, that has feelings. I didn’t think about referencing Boltons excellent lightning talk on Emotions (http://www.developsense.com/2007/05/lightning-talk-on-emotions-and-oracles.html)
There was also a questioning regarding Agile development contradicting my dystopia; wasn’t that a way towards testers being more integrated in software development. I agreed, but also said that if the testers are brought in, in order to create unit tests instead of the developers, maybe it’s not totally positive. We need to look at many aspects.
Another good input was that testers who changes projects and companies, get a more holistic view, since they see more different things.
That’s a valid objection, and my defense was the team aspect, which is lost if people switch companies often. Best of both worlds might be to spend some time at a different place, and then come back.
Questions are the best part of presenting your ideas.

After this I went to a “Manifest Workshop”: Can the Past tell us the Future? This was nice because you got to spend some time talking to people about what has happened with our profession over the years. There are a lot of agreement and good ideas, but you also hear some things you can’t agree with, e.g. “Testers should have the mandate to change”, “Testers must have a special education with a title” etc.
This Manifesto extended over several workshops, and will be available at www.softwaretestingmanifesto.org; but I didn’t sign the acknowledgement list, even though they had an idea of “Anti-Manifesto” statements.
Wednesday ended with a casual session that generated some laughs: The EuroSTAR Testing Quiz.
Erik van Veenendal started Thursday with a keynote on TMMi. This is not my favorite subject, since “process-only” doesn’t say anything about how good the content is.
So I was glad that Erik started with saying that process isn’t enough; you need skilled people and a good infrastructure as well.
I also liked his statement that if you can take a Test Policy and apply it on a similar company; then the Policy isn’t worth anything. 
To succeed with a Process Improvement you need to have Management on board; and unfortunately they only listen to numbers (how long will that be true?)
You also need to spend a lot more time than Friday afternoons (Erik said 60% was necessary); and you need to focus on why you are doing it; what isn’t working good enough?
But Erik sometimes goes back to the Dutch formalized approach: the testing lifecycle should be Test Plan -> Test Design -> Test Script -> Execution.
“People want to change, but they don’t want to be changed.”
It was a good presentation, but unfortunately Test Process Improvement schemes aren’t generally good; most of all because it lacks aspects like creativity, and the ability to find all important errors.
I didn’t ask if there is a risk of becoming too mature, over-ripe, rotten.
On the other hand; you can read TPI, TMMi etc. and get really good ideas on what to do.

Egbert Bouman had presented his five favorite inspirational management Gurus, and how they could be applied to testing.
E.g. “Make company and personal interests run in parallel”, “People are complex.”
But I think it might be dangerous to transfer ideas from different areas without critical thinking. For instance, Collins Hedgehog Concept is about a company in the market that stick to doing what they are great at.
But in testing, we aren’t market-driven, we rather provide services/information to the project; and wouldn’t it be a bad idea to test software in the same way year after year?
He agreed a bit on this at the Gala Award Dinner, but also said that it differs if you are employed by a company, or if you are a consultant. But we discovered that we both were fans of Kierkegaard, and Morten Hougaard joined in this appraisal. “The truth is the subjectivity.”

I skipped Stuart Reid’s workshop on ISO29119. I was interested in how they are working, but I was sure I wouldn’t like the result; too much focus on standardization is killing creativity.
Tim Koomen jumped in at last minute and gave an overview of the testing industry in the past, present and future. He has a good understanding, and talked a bit about Pairwise (I don’t like it), Exploratory (I like it) and Model-based Testing (Surely good in the very few situations it is truly applicable.)
He also had survey results showing that only 1/3 of software testers are using Test Design Techniques (on the other hand; who isn’t implicitly using Equivalence Partitioning all the time?)
“Let’s use the existing techniques a little bit more.”
It was also nice that he brought up the classic Triangle Exercise.

John Kent have created a better Entity Model for Software Testing. “All models are wrong, but some models are useful.”
He used the name Test Condition for Test Ideas, which I think are central in our universe.
He showed a Test Case example, and I really feel sorry for all those testers that has to write or follow Expected Results for each and every test step in a Test Script.
He had a Real World Waterfall Model that started with “Produce requirements which are incomplete and slightly wrong.”
And he ended with the idea that in some situations you might better throw out these Entities and go for only Test Ideas; the tester will know what to do.

After lunch break Michael Bolton gave probably the best testing presentation I have ever seen: Two Futures of Software Testing.
He was wearing horns when talking about the Dark Future that is static and mechanic, and a halo when talking about the Light Future that has the testers human skills in the centre.
His message is very close to my own, with the difference that he has a lot more to say, and he does it extremely well.
It was so good, so I didn’t have time to write down any of all the good things he said, an old version of the presentation is available at http://www.developsense.com/presentations/twofuturesofsoftwaretesting.pdf.

The last keynote was James Lyndsay: Becoming Agile – Reshaping Testing for an Agile Team.
Agile exists, it is growing, and often succeeds with its projects, independent of the inclusion of testers or not!
But the testing performed is confirmatory, and that isn’t enough, so the tester approach is needed.
And in some ways; Agility and Testing is conflicting. Agile is highly optimistic, and testing is perceived as the opposite. Testing is used to be independent, must now be part of a team commitment. To explore for dangerous information might need feel good.
It is not necessary for a tester to know programming to enter an Agile team, but it is an invaluable asset.
He showed some statistics from www.ambysoft.com/surveys/agileFebruary2008.html.
And his main point in a very good presentation was that testers should help the team learn more about what they are building.

The conference was concluded with dinner in Grote Kerk where I sat next to a penetration tester; and beers at a pub, which really encourages many conversations.

Top 3 themes on the conference as I saw it was: People, Complexity and Agile.
EuroSTAR is a good conference, it has a nice aura of knowledge sharing.
Other things: Hoegaarden in Holland has so much more aroma than in Sweden.
Stuart Reid was walking around with a yo-yo.
My main attraction on the EXPO was the company that had a chess board. I played a game, won, and was thereby included in a lottery; and since only one other person won, I had the 50% luck of actually winning an IPhone!
I didn’t have the chance to see: Mieke Gevers, my nice Track Chair; The good title “Imagination is more important than Knowledge”; Fabian Scarano who in his paper and presentation mention serendipity.
It was a tie between really good Mexican and Argentinian dinner.
Isn’t it true that good testing should create low-hanging fruit?

Next Year EuroSTAR will be held in Stockholm, again! A bit boring; Barcelona seems like a more exotic place at least for me.
My current idea is to create a presentation consisting only of images; have a good name like “7 Ways to Explain Software Testing”, and then write a fluffy and nice abstract…
Or maybe try titles like Test Ideas – the Center of the testing universe; Generic Test Ideas for Data Types; What I Learned After 9 Years At Spotfire; The Holistic-Subjective School of Testing; The Eye of a Skilled Software Tester; Verification, Validation, and Everything In Between.
/Rikard