Thoughts on getting Clojure into the enterprise

@odinodin pointed me to @neal4d talk on Introducing Clojure to the enterprise. Having watched it I want to bring forward some of my own thoughts on the matter.

#Guerilla tactics Neal (I hope he’s okay with my using his first name, Mr. Ford sounds so formal, and people might confuse him with Harisson) talks about Cat burglary in his talk, where you let unknown tech piggyback on known tech, as a strategy for getting Clojure in the door. Personally I have greater confidence in some sort of guerilla tactics, that is, find some small project on the side that noone actually cares to much about and implement it in Clojure, and when you’re done (ahead of schedule) you ask for forgiveness.

#The worries of the CTO I’d disagree with Neal in the assumption that the reason the CTO is reluctant to choose Clojure is because she’s scared of being fired. In Norway, it’s impossible to get fired, so it can’t be the reason.

I believe that a good CTO is worried about how to best contribute to her company making money. So, if I were a CTO, I’d definitively rule out a total rewrite of an existing Java application, that’s just foolish. So in order to get Clojure in the door, you need to find an area of your application which is kind of isolated that you can implement new stuff in Clojure.

Also, your CTO will be worried about who else can maintain the code you wrote. When you leave your company to work for Cognitect, who’s going to understand the super-dense code you wrote using zippers and partial functions? Because, you need to remember, she hired the cheapest, not the brightest, developers she could get her hands on (so what are you doing there, anyways?).

Imagine she let’s you start with Clojure, she’ll also have to bring the other devs up to speed on Clojure, and guess what, that costs money. Just because you’re willing to teach yourself Clojure on your spare time, doesn’t mean all your colleagues are willing to do the same.

So the argument you need to be able to make to the CTO is that by adopting Clojure, you and your coworkers will deliver value so much faster that it actually makes up for the investment she has to make to get your team up to speed in Clojure.

Which brings me to the need for

#Success stories I guess more stuff like this from Jay Fields is needed (even though it scares the hell out of me), but I also need the facts to show that a team using Clojure delivers more software cheaper than a similar team using Java.