In agile and especially DevOps approaches the motto is: automated everything! Companies like Facebook claim they do not have testers at all. Microsoft only has SDET (software development engineers in Test), other companies are T-shaping developers to do the testing. New kid on the block is AI and machine learning, that will definitely replace testing I hear people claim. What is really happening globally? Do we no longer need testers? Can we actually automate everything? How can we make valuable software for our clients?
In this presentation I will address questions like:
* Do we need testing? And if so: why is testing important?
* What is the business case of testing?
* Can developers also test? And if so: do we still need testers?
* How can we create quality software?
3. Research& Development:
Dealing with unknownunknowns
Capture everything upfront vs. building new insights
Customers don’t know what they want
Cope with complexity, confusion, change, new insights and halfanswers
So: We need to learn and to deal with risks!
4. Testingand quality?
Testing informs decisions about quality & risk
Quality isvalue to people who matter
Quality is NOT: conformance to requirements
Quality is NOT: the best product possible
Quality products solve the problem and are “good enough”
6. Business caseof LEARNING?
The question is: do you value learning?And if so: how much are you willingto pay?
If you value it enough,you won'tdiscuss the price...
But are thecosts justified?
Or learn to live withoutit … withallrisks involved.
8. Reduce cost oftesting?
Focus on Testability!
Learnability (*)
learning
(*)Learnability: not the old ISO9126definition
My definition: how easy is it tolearn about the value and risks of a product/solution
10. Learning about testability
10 P’s of Testability
Read more about this here:
Team Guide to Software Testability - Better
software through greater testability
By Ash Winter and Rob Meaney
(http://leanpub.com/softwaretestability)
11. Automate everything?
Automated checks provide binary answers to a question to which you
already know the answer.
We need a lot of checks to speed up development by detecting
regression problems.
Exploring can provide new insights into the product. Only people can
provide this insight.
12. Can everybodytest?
Sure. The question is: how good do you want it to be?
Most people do not like testing at all!And that is why they will never be good at it. Nor do they
have time to learn these skills.
We need smart people with "critical distance” to do skilled testing. People determined to find
problems thatmatter.
Problems asin: ”are there problems thatthreaten the value of the product or the on‐time
successfuldelivery?”
So are we stilltalkingabout testing? Or is it muchbroader?
13. About critical distance
We need a diversity in thinking: different mindsets
Opportunity mindset
Solve problems
Ask howand whenquestions
Problem mindset
Search for problems
Ask what if questions
14. Learning in teams: it’s all about loops!
Create learning loops (Plan-Do-Check-Act) in everything you do
Team collaboration and skills are key: optimize your SDLC and processes
Risks are a whole team responsibility
Test your requirements & assumptions continuously
Help programmers go faster: devand release pipelines, automated checks, codequality, measuring the right
metrics
Test enough (good is good enough): test & automation strategy! Oreven better: an integrated quality strategy
based on risks
Use tooling & automation in your exploration
Mitigate “appropriate” risks by nottesting: monitoring, test in production, etc.
15. Learning in teams: it’s all about loops!
Focus onfast learning in ourteams:
Whole team quality!
So are westilltalkingabout testing?