What I learnt at youDevise

Wednesday 31 October 2012 at 16:31 GMT

Today is my last day at youDevise (or TIM Group, as it's now known). I've loved it here, but it's time to move on and do something a little different. With that in mind, I made a list of all the things I've learnt in my two years here. Here it is:

• How to pair
• Ping-pong
• Driver/navigator
• Board and ball
• What a hedge fund is
• What a hedge fund definitely is not
• How an ORM works
• Dependency inversion
• Dependency injection
• Unit testing
• Integration testing
• Browser testing
• Acceptance testing
• Database testing
• Why mocking is actually useful
• Reflection voodoo of the highest order
• The builder pattern
• Build environments
• Continuous integration
• Pipelined builds
• Deployment
• Beer is delicious
• Why you shouldn't think "Engine!" when doing calculations
• Why you should never aim to build an "engine"
• Why frameworks suck
• Why home-rolled frameworks can suck even more
• Why home-rolled frameworks can be excellent
• That <table>s aren't always evil
• Why you shouldn't build your own table generators
• The importance of a decent type system
• How wrapping types in Java makes everything harder
• How making everything strings makes everything even worse
• That waking up in the morning is harder than it looks
• Which conferences I should check out
• Why I should go to a conference
• How to learn at conferences
• How easy it is to try and fix everything at once
• The importance of committing regularly
• The power of the revert command
• How hard it is to make yourself revert
• How gorgeous JavaScript can be
• How terrifying JavaScript can be
• That I really like having a more static type system
• Why a static type system sucks for experimentation
• Trade ideas
• That trade ideas aren't trades
• That if you have a great idea, you should totally do that
• The importance of a domain model
• Why your domain depends on who you're talking to
• Why you shouldn't use your database model as your domain model
• Why ORMs suck
• How to use an ORM in a way that doesn't eat your database
• Why you should just avoid the problem altogether
• Why frameworks REALLY suck
• Why reinventing the wheel never pays off in production
• How reinventing the wheel always pays off in experimentation
• That when Steve Freeman tells you it's a spike, it's a spike
• How to host a developer event
• How to run a developer event
• How to sound enthusiastic even when you're exhausted
• How to ignore the one muppet at the back
• The importance of milling about and chatting
• How to write ten lines of code when one functional line would do just fine
• Why functional programming isn't always the best solution
• How to write terrifying functional code in Java
• Why Java 8 will solve all my problems
• Exactly how hard it is to understand monads
• Exactly how hard it is to explain monads
• Why having free sweets at work is a terrible idea
• Why having free fruit at work is an excellent idea
• The importance of a morning coffee run
• That you should know why you're breaking the rules
• Why Selenium is broken
• How WebDriver is better
• IE6 quirks I never knew existed
• That IE8 is no better
• How hard it is to get a stable cluster of test nodes
• The different concerns that infra have
• How awful Puppet is
• How amazing Puppet is
• That idiomatic Ruby code can still be terrible
• That idiomatic Java code is almost always terrible
• Beer is so goddamn delicious
• That god objects can manifest themselves in all manner of forms
• That you shouldn't try and rewrite them
• That phasing out works pretty well
• That you should fucking COMMIT or REVERT

There we are. Almost a hundred things I learnt. I came up with that list in fifteen minutes on the train last week. I hope that after a couple of years at my next job at Palantir, I can do even better than that.