How to enable organizations to go faster - Fred George (https://www.youtube.com/watch?v=WKTtnVb83mQ)
one hacker away someone else talking about process like my programmer anarchy
simple: eg callcenter complex: eg payroll, taxes complex: advertising, day trading, recommendation engines, should I loan you money
) orgs good at complicated, are bad at complex
) the complicated problems have been solved already
) the money is in complex these days
) usually, you don’t know which problem you are solving you need to see how they behave
) people have a prejudice Trump sees everything as simple and solves them that way Fred George likes complex
how fast can you go?
- technology inhibitors
- process inhibitors
- organisation inhibitors
1) technology inhibitors
1.1) Valley Tech: bigger companies are afraid of using new open source tech
1.2) hardware lead times
-> this kills capacity planning
go buy extra
ops guys are out of a job
1.3) Services impact
result of previous curves
1.4) Incremental Applications
1.5) databases
eventbus becomes the operational database
yes: redundant data + no consistency
but excellent architecture for complex problems
) open source & docker
) functional programming
kafka = scala
rabbitmq = erlang
2) Process Inhibitors
) which type of problem? (conefyn) => there are no experts in complex problems
) what to build complicated, we pushed the customer away complex: try ideas fast
experimentation drives innovation
) rethink interactions
requirements
change the level of interaction with customer (from task to story, from story to feature) what are you trying to achieve
) measure what matters
people will game those metrics
tell programmers what the game is, and they will in
3) Organisational Inhibitors
) overspecialised
-> overhead of communication between specialist
fix the titles
if you want to work on the new projects, you need one of the new titles
rip out desks, and install tables
no leader as title
the person with the highest skill in the current problem, leads the team
based on social anthropology
villages < 100 ppl, never had a fulltime leader
don’t shuffle the teams every new project
) more than just microservices
=> hockeystick on hockeystick
