If you've never heard of CodeRetreat, you're not alone. I first learned of it a few months ago when Patrick Welsh organized one at our workplace.
CodeRetreat is the brainchild of Patrick and Corey Haines. Patrick is an Agile Developer/Coach, follower of strange dietary practices, and all-around great guy. Corey is the hippie* guru of software development in the Midwest, and a top-tier developer.
* I have no idea what Corey's political and religious beliefs are, if any. He just has long hair.
The event was hosted at LeanDog in Cleveland, in their office on a boat. I can't think of a cooler venue or a better bunch of people to learn with and from.
After meeting the team (and the dogs, who are anything but lean) we went out for drinks and discussion. Our host was Jeff "Cheezy" Morgan, who runs the company with a couple of partners who weren't present for the event, but are probably pretty cool as well. Everyone had something interesting to say, especially Paul Nelson (a Windows-hating Mac fanboy, but otherwise excellent pairing and drinking partner).
We used Conway's Game of Life as the focus out our programming work, but made very little progress on solving the game as a problem. Instead, the focus was on learning new (and better) ways to create an emergent design through Test Driven Development. Patrick did a great job of talking us all down from the wall when it came time to delete our code and start over again - something we did every 45 minutes or so.
I had several great conversations while pairing with different people. Some of the things that stuck with me the most came from Corey. Here is a summary of what I wrote down at the time:
The wisdom of Corey:
- You don't change code under red - you add (& duplicate if necessary), then collapse & clean under green.
- You don't write Ruby code. You write a DSL to solve the business problems you have, then implement it using Ruby. If your business logic looks like Ruby, you're doing it wrong. Look at Rspec - it's a DSL, not "Ruby Code".
- "Writers write, always".
- Not sure if what you want to write will work? Write it.
- Doesn't work? Delete it.
- Kinda works? Keep what works, delete the rest.
- Not sure if what you want to write will work? Write it.
I'll post more on this as time goes by...
