My current primary project is a Rails app that depends on PostgreSQL and Redis. I used to run those services directly on my Mac OS development workstation, but didn’t like the untidiness of having services I wasn’t using all the time running all the time. Or that the default install might leave the services listening on all interfaces so that everyone in the coffee shop could poke at my test data. (Nah. I’ve got a firewall and so should you.) I figured, “Hey, I could use Docker to cordon these services off.”

I was a long time user of RVM for installing and switching Rubies. It made my life pretty easy even as I listened to others struggle with it. I was a little uncomfortable putting other things like project-specific environment variables in my .rvmrc files. It seemed dirty, but it worked and I rolled with it for many years.

Eventually, I tried something different. I’d been sold on the idea of using multiple, simpler tools together. Here’s the result.

I don’t think I’ve ever written down my experiences on September 11, 2001. Since I broke down in the shower this morning, I think I probably should.

I spoke this week at ChefConf about how we use Chef at my company to produce an installer for our on-premises product. Slides for the talk are up, but not very useful without my rambling that went with them. Instead of making attendees madly write down the resources slide, here are links to the resources and tools I mentioned in the talk.

I spoke at ChefConf 2014 this week. A few folks who used their Google Fu on me said, “That’s not your blog, is it? Nothing has been posted in forever!” I come from a long-ish career in the security and defense industry where we are notorious for not talking about how we do things. It is taking time and effort to break habits of holding all cards close to the chest.

I gave a lightning talk at Steel City Ruby Conf 2012 today on writing a decent README. Instead of putting the slides up somewhere, I thought I would write up a more detailed post of the talk’s points.