14:00 Meaningful metrics
15:15 Visualization for managers
14:00 Too much, too slow, too many
- Start with collecting data
- Start measure and then see if it make sense
- Experience: very useful info in correlation among
- Committers
- Changed code
- Failed tests
- http://www.gapminder.org/ tool for showing information vs. data
- Measure facts, which underline sprint retrospective points. See their progress from sprint to sprint
- … used the law of two feet…
11:15 Pain of testing legacy software
- Should have many more low-level unit tests than high level end-to-end tests: “Cup of coffee metric”
- Test pyramid often flips and becomes ice-cream cone
- Flaky tests are common, but not inevitable
- Failure in a high level test - a missing unit test. Whenever you fix a failing end-to-end test, you should be adding unit tests tooIt will run a few times before end-to-end test completes. Meanwhile you can fail and fix
- 50% of flaky tests unveil true problemsDo not ignore them
- Good tests are repeatable (give the same results), independent, fast and readable (its purpose and failures are easy to understand) – if one do not apply, be practical with others
- Possible strategy for slow regression tests through UI
- Divide into a few flows (“red line” scenarios)
- Start with restoring backups (on release level) and applying DB upgrade scripts
- Continue with scenario vs. independent tests
- It’s worth maintaining tests as it’s worth maintaining codebase (i.e. refactoring)
- Use “canary builds” or add environment test into the build
- Get a threshold for handling “technical backlog” for every sprint
Meeting purpose
16:30 What tools?
- … not much ;-( …
- … used the law of two feet…
10:00 Test Driven Development
- Advertise CITCON
- Promote open spaces form of conferences
- Entertain you for 40 minutes
- On serious note,
- spot interested in the subject people;
- trigger interest in improving testing in the project;
- provoke creation of plan/approach to improve
- TDD makes you think about design
- You have to be a good designer
- Difficult to apply to a badly designed code – have to mock everything
- Your level of confidence in the code means nothing
- Why to pair? 20% of developers face the code
- Motivations
- Write tests because you do not trust others
- Statistically higher paid developers write tests. Should work another way around
- Not to be afraid of refactoring
- Check in and… it works
- Recommended reading
- Growing Object-Oriented Software Guided by Tests
- Working Effectively with Legacy Code
16:30 Risk Management and
Voodoo charms
- Voodoo charm: “Lets keep an eye on it”
- It seems like we often identify problems, agreeing that it is a problem and that we need to monitor it, but nothing is done
- Brake your system down into components and keep on asking “So what?” to go from
- Failure Analysis -> Impact Analysis -> Risk Analysis -> Business Case
- That’s how potential problem get business value
- In other words SPIN selling: Situation Problem Implications Needs analysis
CITCON – open spaces conference
CITCON 2012
- Arranged by Jeffrey Fredrick and Paul Julius
- Free to participate, but people donate
- Participants introduce themselves
- The rules are explained: empty agenda, rooms and law of two feet
- Topics are proposed, voted
- Agenda is getting its shape
- http://www.citconf.com
Budapest, October 19-20