Erlang

And a historical footnote, WhatsApp on FreeBSD.

For a long stretch through the years where WhatsApp scaled to billions of users with a famously small engineering team, the backend ran on FreeBSD with Erlang on top.

After the meta acquisition, the stack changed.

The point that survives the change is the one Rick Reed and Jan Kuhn made publicly at the time.

So they go off and set off to invent their own.

Erlang came out because the computer science lab had to figure out a way to code the next generation of telecom switches.

Ericsson had had some challenges prior to that, and they set out to fix it.

Elixir came two decades later with a different angle.

When Jose tried to figure out how to bring the power of Erlang to other communities, starting but obviously not finishing with the web.

And now there's a third language from someone who did not have a telecoms or Ruby background.

Correct me if I'm wrong, but based on the conversations you and I have had over the past few years, you came from the world of static typing and you'd been doing quite a bit of work with Erlang but missed your types the moment you basically joined the ecosystem and decided to do something about it.

Static types on a runtime where the most famous phrase is, let it crash.

I mean, some people love them.

I mean, some programmers are like, ooh, let's check, let's try Haskell or Erlang or, you know, whatever.

The Rust is the, you know, the latest hotness in high-performance languages.

So much faster than Python.

But I've worked enough with... object-oriented people and done research on object-oriented languages and so on, that it's clear that for them that is a native way of thinking about the world, just as thinking about mostly functional code is my native way of thinking about the world.

And everybody looks at these other things and we can look at them with respect and be like, Well, it's like my, how can you program without gotos? Same way, how can you program without giving me side effects? What does that even mean? Same way the objects folks look at is like, how can you program without even having objects? Then if you're like an Erlang programmer, you're like, how can you program without having amazing lightweight concurrency? So, you know, Paul Graham has this great, you know, what he calls a blub paradox, right? So the blub paradox is, and he very carefully says, there's a language, you have a language called blub, right? And blub is, it's like, it's just about right, right? And you look at all the stuff that's not in blub, and you look at people using languages with stuff that's not in blub, you're like, how can anyone possibly program without those things? And you look at all these languages that are bigger than blub, and you're like, who Who needs any of those things? I've been blubbed for decades without needing any of those things, right? And we've all got our blubs, and I think the moment you realize that and you recognize that each of us has our own blubs, this is not to say there are not mathematical properties, performance properties, and all these other things.

Of course there are, right? But I know I'm preaching to the choir when I say those things.

I think we just as a community have had a harder time recognizing the sort of the emotional human side of that, of PLs.

But the two episodes actually do belong together.

Now, Because there's a second half to running Erlang at that size, the part we're going to talk about today is the part no one actually puts on a conference poster.

And it's all about keeping a code base that large, healthy for years.

We're in the second decade right now.

And doing so with most of the engineers arriving, never having written Erlang before.

And that itself is a problem on its own.

WhatsApp's answer to this was to go in and invest in the language itself, the formatters, the language servers, the type checkers, and to give most of it back to the community, to the whole Erlang ecosystem.

And for most of our career, the tooling was part of the airline we had to always apologize for.

So we learned based on the site traces and the logs from RabbitMQ on how to get the broker back up or fix it, but not have the right actual code.

Nowadays, we have Erlang developers that are working on RabbitMQ as well, so it's not like that we have not care about RabbitMQ.

It's still a very important part of CloudAMQP, and we have people working on it, submitting code, fixing bugs in the core And also some plugins.

But that was one of the reasons.

That sounds great.

I mean, it sounds like you achieved quite a lot there.

I guess we also need to talk a little bit about why you went away from Erlang.

I guess you mentioned that you didn't have any Erlang developers, but was that the only reason, or what are other reasons that you chose a different language for Daven?

Number 2 is C, at 11.55%.

At the time of writing this script, ladder logic was at number 46 with a 0.28% rating, which put it just below Erlang and just above Haskell.

I'm not sure if that means that ladder logic is not as obscure as you thought it was, but whether it means that Erlang and Haskell are in fact more obscure than you thought they were.

Episode Summary In this episode, we covered the early history of computers in industrial control, the early history of PLCs, including how they got their name, who the major brands are, what they look like physically, a basic description of the abstract machine architecture, a very brief look at what a PLC program is like, the scan concept, the main PLC programming languages, the minor PLC programming languages, the relative popularity of each of the programming languages.

In the next episode, we will take a look at one of the early PLCs from the era when they began seeing widespread use.

It's a control flow in most languages.

And then there's languages like Erlang, who that's the way that it works, right? You just kind of like...

It's a whole way of doing self-healing.

And so I think that depending on what type of programming you're doing, this sort of assertive programming approach where I literally throw in and assert... existential thing that should never happen and then have it there.