The second half of my summary and reflections on the TestBash 2018 conference in Sydney. For the first half: TestBash Sydney 2018 Reflection – Part 1
Avoid Sleepwalking to Failure! On Abstractions and Keeping it Real in Software Teams
The main focus of the talk seemed to be on Hypernormalisation, where you are so much a part of the system that you can’t see beyond it. For example, the message of society is that everything is OK, when really it isn’t, and you know it, but perhaps can’t explain why.
This can apply to testers, where we just accept things the way they are and don’t try and make things better, perhaps because we can’t see/imagine how it could be better.
- Hyper-reality – A world that you interact with, that isn’t real (e.g DisneyWorld). This is how we sometimes relate to software testing.
Abstraction #1 – Quality
The definition of quality is different for everyone. James Bach said in 2009 that Quality is dead, this is a mini-hypernormalisation. He didn’t like the way things were going, and was somewhat acceptant of that.
Abstraction #2 – Measurement
Rich Rogers said in his book “Changing times: Quality for humans in a digital age” in 2017 that we assume when we have tested all of our criteria, that we have good quality (but this is wrong!). If a team’s definition of quality or measurement isn’t true to reality, then why persist with it?
Therefore, we don’t have to accept the way things are just because we don’t see how it could be any better.
Personally I found this a really hard talk to follow (as might be evident in my notes above), in between the old news videos of Russia and the large amount/complexity of content. I felt the point of the talk of not just accepting what is being thrown at us in Quality could have been made much simpler and more time spent on applications of how to do this, or what areas of the job we should be actively looking for better ways to do.
I do strongly agree with the idea of not accepting bad practices just because that’s how it is now, and striving to make things better (ie being agile), but I found the delivery missed the mark a little bit.
Test Representatives – An Alternative Approach to Test Practice Management
by Georgia de Pont – @georgia_chunn
Georgia started by giving some of the context of the patterns/techniques used at Tyro, being Pairing XP, TDD and the Spotify model of squads. We then followed the journey that the testing team at Tyro have gone through, starting by moving from a disconnected team into having embedded testers.
This created some challenges:
- Loss of alignment and knowledge sharing within the testing group
- Lack of consistency in test approaches and usage of test members within the team
- Lack of a test manager (it was a flat structure)
The question arose of how to address these issues, should they hire a test manager or take a grassroots approach? (Evidently, they took the grassroots approach).
Each product tribe selected a test representative to act as representative. They would support their tribe’s testers, gather info on issues and be a point of contact for that tribe. The representatives came up with a pipeline of improvement initiatives to work through, with ideas coming in a variety of ways. The representative group, “rep group”, would share the info back to their test teams and would have regular meetings with the TestOps team.
Some of the initiatives:
- Clarity of role of embedded tester, e.g. what are the test practices they can offer
- Improving the recruitment process. Candidates were given a take home test, which wasn’t reflective of the workplace because people usually pair. So they created a pairing exercise instead.
- Onboarding new test engineers and probation expectations
- Performance criteria for test engineers for them and their leads to use
- Upskilling test engineers (an ongoing effort) of making sure there is available training, organising internal and external speakers, conferences, etc.
- Co-ordinating engineering-wide test efforts
- Developing a quality engineering strategy, involving various stakeholders, to identify any current roadblocks in the testing efforts and work to remove them
Some steps for success to make a representative group work in other workplaces:
- Start small. Think about how many people per teams (aim for 1 rep per 2-4 teams)
- In forming the rep group, consider whether best to select people or ask for volunteers. (they selected people to start with)
- Communicate the work of this group to the rest of the organisation
- Maybe hook into existing communications
- Include ways to get involved
- Ask for feedback (e.g. surveys)
- Now they do this more informally
- Asking: Is it working/effective?
- Run Rep group retrospectives
- Ensure support from engineering leadership
- Maybe include them in meetings
- Get their feedback/input
- Give them an awareness of upcoming testing concerns
- Get budget support
Georgia did quite well for her first talk, it was clear that she had practiced presenting this before, she had helpful slides to guide the content and carried confidence in her presentation. With practice presenting she will only get better, and it will be interesting to see her speak again in a few years with more experience.
The content was relevant, describing a smart solution to problems that many organisations are now feeling with the recent direction testing roles are taking. Even for companies not in a position to form a ‘rep group’ there were definite takeaways in how a testing group can, and should, interact with other stakeholders within Engineering. The initiatives the group came up with look really interesting, and I would’ve liked to hear much more about those as there appeared to be some directly applicable opportunities in there, perhaps this will be the subject of future talks?
Enchanting Experiences – The Future of Mobile Apps
Pari finished out the day as the final keynote, speaking on where she sees the design of applications and devices heading in a heavily connected world and what opportunities could be unlocked.
It started with Machine Input, with a definition that this is all about information that devices can find on their own.
Products are developing all the time, getting more complex, with more interfaces, but perhaps what we need is to get simpler again, and automatically detect more of the options we would otherwise have to choose (through machine input).
We then looked at a range of machine input types
- Camera – e.g. auto detect details of a credit card held in front of the camera and pre-enter
- Location – detect where you are and make suggestions, including smart defaults like detecting the country and local currency
- History – Each time you fill out forms, remember what data you put in, particularly between site visits.
- There’s lots of other sensors in our phones that can be hooked into.
Could we detect which floor someone is on, which shop they are in, what they have recently searched for and offer them a discount as they walk around the shop, without prompting?
The internet of things is leading to a more and more connected home and we looked at some ideas companies like Google are developing in this space.
- Designing great products is about great experiences.
- Security and privacy is really important with all this data being available
- Devices should not dictate us and how we live, they should make things easier and work for us
- Learn about design and usability to see how it can impact your testing efforts/plan.
No particular takeaways from this one, as I felt it wasn’t really a talk targeted for testers, more of a talk for developers or designers who can more directly incorporate Pari’s design thoughts into their work. It was interesting to think through some of the options made possible in various parts of our lives through machine learning from the point of view of: “that would be cool to see!”. There will definitely be challenges in testing any of these types of features, which wasn’t covered much besides acknowledging that security will be key with all these features. Also the content spawned thought processes on what things would need to be considered in testing these sort of technologies.
99 Second talks
99 second talks then wrapped up the day with about 20 people speaking. For most people it was a chance to practice public speaking in a somewhat non-threatening environment. For others it was a chance to promote their company/group, or just a chance to introduce a topic/technique. Hard to take much in given the format, and I didn’t have notes of any takeaways from this, but some interesting topics raised.
Overall thoughts on the conference
One of the big selling points leading up to TestBash was the community feel of the event, particularly seen in the pre and post conference meetups. I was unable to attend either, so I didn’t get the full experience and can’t really comment on how the conference lived up to that hype. The conference day felt much like most other conferences I’ve been to in terms of interactions between sessions and general flow of the day.
There wasn’t much chance for questions with most of the talks, which was unfortunate, as there can be some great discussion coming out of this as people dive in to the bits that were missed or are really interesting to them. Though, in saying that, question time can be hit and miss depending on which direction the conversation gets steered towards, so not a big concern.
The talk selection was pretty good considering it was a single track, so every talk had to be chosen for being applicable to a wide audience, covering a generally applicable topic, but trying to make that topic interesting enough for anyone to learn from. This is a pretty hard criteria to meet, so hats off to the selection team. If you were wanting a deep dive on particular topics, this is perhaps not going to be the best type of conference to attend, but it could still happen on the right topic.
Speakers are well looked after by TestBash, in communications and compensation, though as a local speaker, there was nothing to be compensated for, so that is definitely more of a perk for non-local speakers.
The swag was better than most conferences, aiming for practical items that can be re-used beyond the day.
Overall I found the day enjoyable, I got good feedback on my talk and I have some renewed ideas and techniques to take back into my workplace from other talks which makes for a successful conference in my regards. This is helped by sharing my thoughts afterwards, which I recommend others try after any conference/training, even if just to peers in their workplace.
Would I go to the next one? Maybe, if I wasn’t speaking I would need to look at the schedule first, and see what talks are being covered. I’m definitely glad we have another testing conference in Australia, we seem to be adding more to the mix every year at the moment, and each one lifts the overall quality of presentations as people get more practice, we get more international speakers and a wider range of topics and conference setups to choose from. Each conference will have to work harder and harder to stand out and provide a good experience to attendees, which is a win for everyone! So I’m glad TestBash has come to Australia, and is coming back next year!