Late October 2019 I had the chance to visit a conference where I had not yet been before: The Imbus QS-Days. I was curious to understand the difference to other conferences where I had been, for example the Agile Testing Days (ATD). Also, I had the opportunity to give my first own presentation at a testers conference – talking in front of people who are so experienced would be a new experience for me.
Lots of my co-testers warned me of the different atmosphere – and dress code as supposed to the ATD: suits and ties and business costumes instead of casual stuff and fun costumes. And they connected a kind of atmosphere and attitude to it: stiff, no fun. Let me tell you about my own experiences… To begin, the lineup of topics was really great. It definitely had the potential to expand my horizon (“Expanding Horizons” was the title of the conference). As on almost every conference I have read about in 2019, QA people are worried about their future role, coming with AI, IoT, and also other new research on how to test – and there have been talks about those trends. Last but not least we also saw a talk about how agile changes testing – this is definitely something you would think is obvious if and when you listen to agile testers… But the talk by Alon Linetzki (@ALinetzki) was great – and perfect for the audience: not everybody was an agile evangelist at the conference. And here is the largest difference I would see:
Testers who work in large projects and not in agile teams seem not that well represented in Agile Testing Days. In QS-Days we had several talks about large projects. The audience very often came from large companies with strong backend impact – and not fancy front-end. My favorite talks on huge testing complexity was about the test of the cabin intercommunication data systems at Airbus (many testrigs with a multitude of automated test settings) and about how to automatically detect the most relevant tests in a huge interconnected system.
Instead of just using AI or computing power, Teamscale used human intelligence and presented an approach for change driven testing. Starting with a structural presentation of the system under test and its test coverage, the tool will return using which few tests the largest coverage can be achieved. 80% of the code often could be tested with the first 1% of test code – they say. Also, non tested changes can be detected. This especially makes a difference in large systems. Interesting side fact: this company is a university spin off! Big brains – and also very interesting people. Had a great talk with @ElmarJuergens about his products. If I were clever enough – and his company would not be in Munich – I would love to work there! Another spin off of the university of Munich was “qualicon” – a company using text analytics to improve requirements and test cases. @HenningFemmer very nicely explained how this can help if you have a lot of text to be reviewed and thus improved.
End of day one is was about time to give my own presentation. After having heard several very good talks, imposter severely hit me. I was about to not give the talk – but then had the great idea to ask for help (in the slack channel “women in test” in the conference channel). And there came two angels: Emna Ayadi (@emna-ayadi) and Varuna Srivastava (@vibranttester) who calmed me down – and made sure I did not run away…
THANK YOU SO MUCH FOR THAT!
And actually, my talk went very well. It was about how I, as test manager, can help overwhelmed business users being told to test. Even though they do not know how to test and they do not have time they can start to like it. I talked about motivation, learning strategies and change management – my audience liked it. Quite a few of the about 100 listeners came up to me to discuss and exchange further ideas. I was so happy! The talked based on my ideas about “4 t to success” and was enhanced by statistical data I collected during the last year.
No testing conference without talking about AI. I did not go to all talks to this topic, only want to mention three: a fish bowl discussion about “KI in Testing Tools”. I was rather disappointed as from my perspective, most of the discussion was off topic. It was about the question if AI is necessary in testing tools, and not about what AI does in testing tools. Too bad. My next attempt at approaching this topic was a talk by Dr. Qin Liu (Tongji University) about “A Confidence Index Based Test Approach for AI Systems”. She presented a confidence index approach to manage the risk associated with the “black-box” nature of the AI system in order to extend standard risk based testing techniques. Wow. Heavy stuff, I am not sure I got all.
The highlight of the AI talks was the final keynote by Prof. Dr. Katharina Zweig (@nettwerkerin) “Test the bot – Would it pass Ethics 101”. With a few nice examples she explained the morale and dangers of just believing AI. She used the “beer and diaper paradoxon” to showcase how people tend to just believe machines… Even more convincing was the example how discrimination algorithms do need ethics: in case of discrimination of “bad” and “good” guys, you will eventually need to decide if you prefer to risk to let bad guys run free or to convict good people. On my reading list: her book “Ein Algorithmus hat kein Taktgefühl” (An algorithm has no tact).
Coming back to the dress code – yeah, there have been a few suits. But the people definitely have not been stiff at all. And I definitely liked the quality and content of the Imbus QS-days. I learned so much and could have learned more if I had had more time. I absolutely would be looking forwards to go there again. In contrast, during Agile Testing Days I preferred the extremely interesting discussions and contacts allover the conference. Also, way more topics on human and social interaction could be found. Sorry to say, the party at ATD definitely is way more awesome and never reached!! So I guess both conferences are worth visiting. I love it and hope to meet all of you again!