Elm. Rust. Spark. TensorFlow. R. Haskell. Mesos. AR. Diggory.

Got you. ‘Diggory’ is a character from Harry Potter, not a technology. Or at least it wasn’t one at the time of writing. By the time you read this abstract, Diggory will probably be a JS build system.

If you read Hacker News on a regular basis, you might suffer from a combined sense of despair and impostor syndrome. How is it that everyone else’s stack sounds so much cooler than yours? How are you supposed to keep track of everything? How do you know when it’s time to adopt one of these things that sound like a lesser-known Muppet? How will you ever get another job without all of these things on your resume?

There’s a lot to be said for using tried technology, but that brings a different set of problems. Engineers get frustrated when they never get to play on the bleeding edge. Sometimes you miss a trend and then you find yourself with a million lines of code that you can’t hire anyone to work on, with no tests, because that was another trend you missed.

When should you adopt a new technology? How do you know when to steer clear? When should you throw away the old? And just how many different platforms should you support at one time?

All these questions could be lost in time, like tears in rain. Or we could try to answer them in the course of this talk.