At HCA Healthcare, we built Waterpark– a continuously available, geographically distributed, high velocity Enterprise Integration Platform. We built it from scratch in Elixir. Elixir and the ErlangVM are very powerful tools. Need to model millions of things? No problem. Need processes to self-heal when things break? No problem. Need to model millions of things that self heal on computers scattered across a continent? Hmmm… for that you’ve got some work to do, but in Elixir that work is very doable and fun. In this nuts-and-bolts session, we will discuss the distributed computing bits that are in-the-box with Elixir and what is missing. We will cover the missing bits that we built and “why”. Topics include: hashing functions, distributed process registries, the mailroom pattern, server topology, leader election, consensus, actor replication and recovery, and hot code loading. 

At HCA Healthcare, we built Waterpark– a continuously available, geographically distributed, high velocity Enterprise Integration Platform. We built it from scratch in Elixir. Elixir and the ErlangVM are very powerful tools. Need to model millions of things? No problem. Need processes to self-heal when things break? No problem. Need to model millions of things that self heal on computers scattered across a continent? Hmmm… for that you’ve got some work to do, but in Elixir that work is very doable and fun. In this nuts-and-bolts session, we will discuss the distributed computing bits that are in-the-box with Elixir and what is missing. We will cover the missing bits that we built and “why”. Topics include: hashing functions, distributed process registries, the mailroom pattern, server topology, leader election, consensus, actor replication and recovery, and hot code loading. 

At HCA Healthcare, we built Waterpark– a continuously available, geographically distributed, high velocity Enterprise Integration Platform. We built it from scratch in Elixir. Elixir and the ErlangVM are very powerful tools. Need to model millions of things? No problem. Need processes to self-heal when things break? No problem. Need to model millions of things that self heal on computers scattered across a continent? Hmmm… for that you’ve got some work to do, but in Elixir that work is very doable and fun. In this nuts-and-bolts session, we will discuss the distributed computing bits that are in-the-box with Elixir and what is missing. We will cover the missing bits that we built and “why”. Topics include: hashing functions, distributed process registries, the mailroom pattern, server topology, leader election, consensus, actor replication and recovery, and hot code loading. 

"Poka yoke” (po-kah yo-kay) is a term from Japanese manufacturing that means "mistake proofing". It is a powerful concept (equally powerful in software) in which the design of a system shapes usage so that human error is prevented or corrected before it can cause real harm. In this session we will explore how and why FP maps so well onto the poka yoke concept. We will discuss the merits of mistake proofing over diligence, and rethink our approaches to craft and quality. You will leave with a new lens to evaluate languages, tools, and your own code. Examples in Elixir, Elm, F#, and Idris.

Elixir makes me happy, and I bet it will make you happy too. OK, but why is that? Elixir is a modern functional programming language with productive “dev joy” tooling, and it sits atop a decades-stable, crazy-scale, battle-tested foundation: the ErlangVM and OTP. In under three years after Elixir reached 1.0 it has already topped big, popular languages like C# on Stack Overflow’s "2017 Most Loved Language” list. After this session you will understand what makes Elixir special, the ecosystem will be illuminated, and you will have the "why, what, and how" knowledge to confidently begin your Elixir journey.