Hey everyone,
The elitism of Lispers is legendary and predates Clojure by a lot. The phenomenon has been given a name: Smug Lisp Weenie.
So this is just my opinion, but I think Clojure is having trouble “crossing the chasm”. There’s a book by that name that describes a theory of how products make it to the mainstream. The theory is called “Diffusion of innovations”. Clojure is a well-designed language and so has a lot of natural appeal. It easily got the first 16% of the “market” with innovators and early adopters. Rich Hickey’s talks, a lack of other options early in its life, and a lack of “practical” Lisps gave Clojure a huge, early migration of people ready to learn the warts and brave the lack of tooling.
The thing is, those people run out. If you want to get to that giant hump where most people are, you have to really appeal to the more conservative decision-makers. The early majority are looking for practical evidence that this is a good decision for them. They’re not concerned with features, and they won’t ever learn the language just to find out all the wonderful things we love about the language. They may already be interested in Clojure, but they’re waiting for a critical mass of their peers to jump in, too. It’s a chicken and egg problem. I believe this is where the excuse of fear of not being able to hire comes from. You can read more here.
The biggest issue is that Clojure programmers are all early adopters. We can’t sell Clojure to the majority with the same ideas that attracted us. And the only way to figure out what will convince others is to hit the pavement and talk to people who are kind of interested in Clojure but have not adopted it for some reason. Empathize with them, figure out what they want, and then give it to them.
About types: seeing all of the flamewars online, I try to elevate the conversation and transcend the differences most people fight over. That doesn’t always help others or the discussion, but it often helps me. It’s therapeutic and helps me develop my views of things. Another way it helps is to avoid getting caught in the repetitive back-and-forth of the flame trench warfare. I guess I’m like a war journalist on the Maginot line.
As an example, I wrote a post comparing the perspectives of static and dynamic typists. When I diagrammed it, I realized that we have fundamentally opposing assumptions. Static typists believe that untyped programs are a subset of typed programs, probably because that’s how you would implement a dynamic solution in a static one. On the other hand, dynamic typists see typed programs as a subset of dynamic–they’re programs checked for type correctness.
The reason I like this is every time I see a discussion, it’s very clear where the arguments come from. It’s in the nature of this type of discussion–the ones that get nowhere but people are fighting very hard–is that people are talking past each other. It is so clear when someone is stuck in either Church or Curry thinking. When they’re stuck, I have very little desire to argue with them.
Rock on!
Eric