A Workshop on Testing for Programmers

An idea had been twirling around my brain for a while. Then it became more and more insistent on breaking into the realm of reality. In the context of continuously learning about testing, and in the context of recognizing the ongoing challenges of programmers and testers working together, I realized that I need to do more about it. Quick conversations over lunch or in the hallway about one or other aspect about testing just didn’t seem to be enough. There needed to be something that’s a full package, coherent,  concentrated and experiential. There needed to be something that would renew the foundation for future conversations between testers and programmers.

Mission: create a full-day workshop on testing that would be tailored and targeted at the programmers I work with.

I consider this workshop as one of the great outcomes of Transpection Tuesdays. Without the work I’ve done with Erik Brickarp and without his support and encouragement, it would’ve taken me much longer to muster up the confidence to create and carry out the workshop.

The keyword here is “tailored”. Fun and exercises aside, I focused on tying together the general introduction into testing, the way testing had been done by my team, and whatever else was relevant in the company’s context in relation to testing. I wanted to frame some of the specific problems in our company as testing problems and have a discussion about them.

Goals:

  • give a concentrated introduction to testing so that fragments of what programmers already knew were tied into a whole
  • introduce and explain testing problems to  programmers so that they could be understood and discussed better
  • immerse programmers in some testing problems so that they would see and experience them up close (followed by a discussion and explanations)
  • explain what testers do and tie it into their experiences of working with testers so that programmers can make sense of their experience and forge even better relationships with testers
  • explain the tester’s mindset and vocabulary so that further conversations between testers and programmers could be more productive

Why I’m Sharing This

I’m sharing this just because I believe from my experience that creating such a workshop can be useful and educational in many ways, and there are surely people out there who need just a bit of nudging to do something similar where they work.

Slides and some accompanying materials can be found here.

It doesn’t contain all handouts because well… what you think you need to provide to people you work with will really depend on the kind of workshop you create. I also didn’t include my notes for slides because they have quite a lot of context-specific stuff in them and it didn’t make sense to try to clean it up.

You can grab them and run but it would be so much more useful for you if you read the following as well 🙂 Then you will know WHY it is the way it is.

Planning: content and structure

Preliminary notes

One of my favourite ways of getting some mental work done is to go to a long lunch alone, take my notebook with me and just write my ideas down. I did quite a bit of work on the workshop in that manner, clarifying my initial ideas about the structure of the workshop, taking notes about what kind of exercises to use etc.

Mindmapping

  • I created a mindmap using Xmind to sketch the initial structure of the workshop. The mindmap I’m sharing with you isn’t perfect and it looks sketchy. I just used it as a starting point.
  • I mindmapped the big sections and also drafted ideas for subtopics to be covered in sections
  • I added labels to each chunk estimating the time I’d be spending on it. Then I summed them up and the result was the time I was going to spend on each section. And from there I could calculate the time I’d spend on the whole workshop.
  • Estimating the time spent helped me balance the sections and decide whether to include or exclude content. It also kept me in check and guarded against adding too much.

Slides and notes

  • Based on the mindmap structure, I started drafting my PowerPoint slides, taking notes about the content and gathering relevant references.
  • I kept track of references in a separate document to make sure I can hand that list to the attendees later.
  • Working with the mindmap and slides I did a lot of gauging of the flow and used internal monologue or tried explaining some concepts to myself. The purpose was to make sure the flow of ideas made sense and that each section built upon the previous one. In that process I made structural changes to the mindmap which provided me with a good backbone and overview of the workshop.
  • For each section, I added a separate title slide and a recap slide. The idea was that the day will be so packed with information that it’s better to add some points where I can concisely summarize what we just talked about. All those recaps would then go on one sheet of paper. If that was all they ever kept from the workshop… they would have the essential ideas captured.

Exercises

General idea behind beer

  • I wanted to have a “thin red line” that would run through the workshop. There would be so much ground covered in terms of concepts that I felt like falling back on a single constant thread would help pull us all back in.
  • I found that thin red line in the workshop by Louise Perold on test planning that I attended at Copenhagen Context this January. She used bottles of beer as the product in the workshop and it worked really well for a test planning exercise. Sure, it’s not software but… beer is fun, fairly well understood and doesn’t need much explanation. So I discussed this with Louise and she was fine with me picking up the idea and using it myself. Many thanks for that! 🙂

Kick-off exercise

  • Testing the beer was to be the thin red line and the kick-off exercise.  Each team of 3-4 people got two different bottles of beer. The only brief I would give to testing beer at 10am in the morning was to “test the beer and explore it” in 20 minutes. I didn’t give them the bottle opener as I wanted to see if any would ask for it. I gave them a sheet of paper and colorful markers to write down whatever they wanted about the beer they tested. I made observations and took notes about what they said in their interactions while they were testing the beer – I’d use these notes for debriefing and for tying the discussed topics with the rest of the workshop. For instance, if I’d ask how they knew the beer was OK, they could give me answers like “from my personal experience” and later when getting to oracles, we could discuss their initial test run in the light of heuristics and oracles (or anything else we covered during the workshop).

Follow-up exercises with beer

  • Testing mission – since very few asked why they were testing the beer and what was supposed to be accomplished, I used this as an example of how not to approach testing… and then asked them to figure out a few testing missions for beer that would frame the testing activities. I also brought in the project lifecycle to show how different things can be tested for over time.
  • Oracles – since it wasn’t that easy to find actual issues with the beer, it gave me a good segue into asking why couldn’t they spot some problems. And from there we could get into oracles and how they’re useful for testing. Later I used as another recap of the kick-off exercise when trying to come with some oracles for testing the beer and pointing out the ones they had actually used but didn’t realize it.
  • Test strategy – I gave them a brief, then asked to come up with a simple description of a strategy. I mostly meant this one as the “dip into this but don’t get too deep” kind of an exercise. The point was to teach about test strategy per se but to make them feel how difficult it may be to word it. Also, I gave them the whole HTSM on paper but limited what they had to use. It was probably a bit too much…. in terms of balancing time and materials. But I feel like they got the sense of complexity out of it at least…
  • Creating a script for opening the beer. This was a pretty fun one… The idea was to create an experience where you’re supposed to do something simple – just follow the instructions. But then describing how to open a beer in a specific manner isn’t THAT easy. So on the one hand, people could get creative (some drew up pictures to illustrate the steps) but then when executing the steps, I limited them in their interpretation. One of the coolest instances was when the script was followed precisely but the beer bottle ended up with its cap on not off  🙂 There was a very interesting contextual cue that a programmer decided to “misinterpret”…
  • Testability – when discussing testability, we returned to the beer trying figure out what would make it easier to test beer.

Other logistics

  • I divided programmers into 3 groups of 9 which meant I ran this workshop 3 times. This seemed like a manageable amount of people for me.
  • Since I knew them all well (I’ve worked with them for years), I tried to balance out the groups so that programmers from different teams could interact as well.
  • I tried to make the first group the most difficult for me in order to make sure I will find any issues in how I’ve set up the workshop and fix them before the next one. That one I got right – they were active, wanted to argue (not just with me… even more amongst themselves), wanted to discuss problems and potential solutions to them. I was able to figure out some issues in the order of sections which was a good thing to fix.
  • I had snacks ordered for the breaks and pizza for lunch – just to make it easier for everyone.

Lessons learned

  •  Oh my god… was I tired after the first workshop! I felt completely exhausted. It’s the type of feeling when everything is in slow motion… like reaaaaaallllyy slllooowwww…. But I still felt great about how it all turned out. Luckily, the 2nd and 3rd runs weren’t as exhausting.
  • I learned how much time actually goes into putting a workshop together. I think I may have spent at least 100 hours when preparing for it.
  • I learned that I enjoy working on a workshop like this. It helped me to freshen up what I know (well, it’s fairly well known that teaching someone teaches yourself a lot, too). I also enjoyed delivering workshop even though I got tired and some people started arguing a lot 🙂 It was fun to observe how programmers from the same team started arguing how to test better and what they could or couldn’t do on their end. I felt like getting such a conversation going was an accomplishment on its own.
  • I wasn’t very good at using all of the information I picked up from eavesdropping on the conversations taking place during the exercises. I guess this is a matter of practice and a few other things… like having a much better command of the material. I feel like I could’ve done this part a bit better.
  • The second half of the workshop could’ve used more actively engaging exercises. For some reason I figured people could be too tired for them… but I guess there should’ve been something that would shake people up 🙂
  • After the first run, I did change the order of topics in part two: I figured I should introduce the terminology of oracles and heuristics before starting to talk about heuristic test strategy model. Else I would introduce words I would only explain later.
  • Good feedback from participants was along the lines of “I had no idea testing is so complex” and “I didn’t know how much more there was to it”. Someone even said that “I figured testers had to be smarter than programmers” 🙂 Since I left the company about a month after wrapping up the workshop series, I can’t tell what kind of overall influence the workshop had. But a few programmers I’ve talked to afterwards have told me they’ve thought about the workshop and what they took from it, so… I guess that’s a good thing. Even though I didn’t mean it this way, the workshops turned out to be my gift to my colleagues before I left.
  • It was also pointed out the using the image of a house helped to pull together some loose ends. Talking about the point of a house (it’s where people live in) nicely summed up the reason we ever develop something – there is a purpose for some human beings.
  • Would I change anything next time? Yes. I would probably pick a piece of software or prepare it myself (you know, learning challenges for me too…). I’m not sure I’d pick a different one for each exercise or not but I’d consider it.

 

Am I happy with my first attempt at putting together a workshop? Yes, quite 🙂

 

Cheers!

Some Notes from Copenhagen Context 2014

Copenhagen Context is a new conference held in… Copenhagen, obviously. The focus was on longer keynote-style talks, so the whole conference was comprised of 5 talks. However, some of them introduced practical exercises as well. I’d go so far as to say that such talks align with the spirit of the context-driven community, i.e. ideas you introduced should be applied in practice.

Overall, I think there was a good balance between philosophical, inspirational and practical talks covering different aspects of context-driven testing. I left feeling content and equipped with useful material. Thanks to Morten Hougaard (and Anette) who put such a great program together!

Michael Bolton took his time to get to the whisky part. I almost got thirsty. He talked about the principles of context-driven school and his preferred order. I agree with putting the third one (people, working together, are the most important part of any project’s context) to the first place because it goes well with my own observations.

He also talked about the paradox of CDT: instead of being context-driven we sometimes have to be context-driving as we need to change the context to persuade someone that there is no one true way. That’s when you’re not context-driven.

Though, when I think about it an tie it with my experience, I end up with something like “the context can drive me to become a driver, that is due to something in the context I decide it has to be changed”…

Carsten Feilberg encouraged us to use child labor for testing remember how we used to be natural explorers as children, and how heuristics help us comes up with test ideas. It was useful to be reminded to go meta yourself and recognize the heuristics you actually use. I guess that if I tried, I’d initially come up with fairly context-specific heuristics but I’d like to try and generalize them. Although I think there’s a fairly good chance I’ll just re-discover what others have put out there. It would be a good awareness exercise anyway.

Carsten didn’t really trust a room full of testers with his remote-controlled helicopter (the remote was out of our reach but we could gently touch the helicopter at least). But it was fun to think about test ideas for it. Funnily enough, my first couple of ideas were pretty much a perfect match to the mission of Carsten’s test session of the helicopter: finding out the range of the remote control and if the radio signal could get obstructed by objects, and if the helicopter could get lost (apparently, my hunch was right and Carsten did find a bug with that :)).

I missed out on Louise Perold‘s beer testing session at Let’s Test last year but I was in luck this time in Copenhagen. Many interesting conversations were sparked and cool ideas introduced as people were thinking of testing a bottle of beer. I really think it is an elegant example to use beer as the product to test because it’s versatile, informal, and pretty much everyone can relate to it in some way.

I really liked the idea of “talking to the function in your own words”. I tend to forget the power of paraphrasing from time to time but it is really powerful for clarifying my own thinking as well as detecting fuzziness in other people’s thoughts.

I’m going to work through the slides and the ideas again but there’s a lot of stuff I want to share with my team.

Huib Schoots‘ talk was about what he believes are the key ingredients of becoming a great GREAT tester: passion, learning, and courage. He emphasized practicing testing and asking feedback for learning purposes. Since Huib challenged the room of testers on knowing the test techniques they use (or rather… the number of them), it made me think of what are the criteria for professionalism in context-driven community. I asked him on Twitter, so now I’m looking for that blog post 🙂 Though I’m thinking now that “heuristics for recognizing professionalism” is a better way of putting it (inspired by “Context-Driven (presentation) heuristics“).

Finally, Henrik Emilsson encouraged us to put names and faces next to stakeholder roles, think of them as people (and the feelings we have towards them), and THEN think of how to communicate the test strategy to that person. The session included to practical exercises which I liked (they focused on writing) but I shall try this out on my own. I gave it my best but I felt I was getting tired already.

One of the most valuable thoughts for me from Henrik’s talk was related to having invested or detached stakeholders. I’m wondering if great test strategy communication could be way to gently “re-attach” stakeholders to projects. And I can see how this goes well with systems thinking when solving problems… So yes, a lot of good stuff to think about for me as I’ve been trying to improve my communication of test results over the past 6 months (still have to blog about it…).

And then I met my testing friends from Twitter and had a great time at a pub solving a testing challenge. There was other stuff but… what happens in Copenhagen, stays in Copenhagen 🙂

See you there next year?