Hold up! New talk incoming!

I was going be at Øredev this November. Now I'm going to double be there. I recommend poking your employer to get you tickets and send you there, get the latest trends and make the good connections. I was there last year and had a great time. If you go to my talk I will personally tell you how to do Embedded Linux better than everyone else is doing it.

Also have it on good authority that Fly and by extension Tigris are likely to show up. Hope to have some fun with them.

It is kind of a perfect time of the year in Sweden to be in-doors and nerding out because the weather will be terrible and it will be a long time until we see the sun.

Note: I am being compensated with tickets to the event and accomodations to help promote the event. The last one I was at was a very good event so I bothered to maintain the relationship and make something happen this year as well.

Pick the slimmest thing you can get away with

Niche down. Always. You start wide. That's the default. Then you niche down.

This is a suggestion, not a law. We all by necessity build a general  skillset and you have to start general because there is a lot to absorb and internalize in working with code and computers. Specialists are not that common beyond people that stay in their lane as a Java or .Net dev. That's because those are not just ecosystems but entire worlds unto themselves. You can live there forever and never travel the world if you like.

I started a company shortly before formally incorporating Underjord and while I was hoping to try stuff with Elixir I thought about morePython. I thought about getting good at AWS. I considered specializing in maintaining legacy systems. I don't think I have the temperament for either of those. And then I ended up falling into Elixir more or less. But I was casting about for direction. It wasn't certainty until I had done at least two Elixir clients and it felt repeatable.

It was in the last year I saw that I could probably get enough work for me and colleagues still while focusing primarily on Nerves and at least moving us in that direction. Focusing on that ecosystem.

But I didn't stop the thinking there. The plan wasn't just to do Nerves. My plan was to focus on security and Nerves. Elixir is considered niche, though it is quite wide. Nerves is a niche within that niche and is actually a somewhat limited pool though I keep finding more and more of it every week. Setting a security lens on top of that makes it a needlepoint. As an aside the plan didn't hold, I now do a startup instead. I think it would have worked though and I would have pursued it.

Why focus so tightly? To make the answer obvious on the other end. Put a Nerves-specialist next to a Nerves security specialist. The question becomes, do you have security concerns? Probably. Then the security specialist is preferable. If what you need is UI work and there is someone running a Nerves UI consultancy (arguably Formrausch) they'll beat my specialization. If what you need is primarily board bring-up or porting a Yocto-supported board to Nerves you should actually talk to Redwire, though you might not know it yet because that's a hard specialization to establish in one word. Every Nerves agency or consultant I've talked to (and there are more than you'd think) can do the basics of building out a reasonable application on Nerves, customize a system, build some web or UI.

I'm also a marketer and security just has that whole hacker romanticism around it which with a technical crowd tends to make people assume you are two points better in every stat than they'd assume about anyone else. Or maybe that's just me. Anyway, I've always enjoyed security stuff so that was aspirational.

It is fairly rare that a specialization excludes the broader competency. You may get a UI designer that can't do frontend because visual design is not a specialization of frontend development. You may get a database guru that doesn't write any code beyond SQL. Because databases can be their own field distinct from programming. But if you find a Rust developer specialized in zero-copy data pipelines using the Arrow ecosystem. They'll also know how to program other things in Rust. You may find an embedded Linux developer that can't do web. Because that's not a pre-requisite to that specialization.

Going for a needlepoint makes you stick out. If you go to someone in the Elixir ecosystem and go "okay, I have an Elixir project I need built, where can I turn?" there will be a smattering of consultancies that come up. DockYard, Alembic, SmartLogic, Mimiquate and if they know the space a bit better they'll rank Erlang Solutions. If the question is Erlang project, instead Erlang Solutions is the first answer. If the question is Nerves, there are fewer that have good answers but the answers will be more precise. Since I'm loud, many people would point at me. I'd pass people on to Redwire as a default since I only take on very small very precise things right now.

So Elixir inside-baseball aside. What build requirement is your skillset the easy answer to?

What would be your speciality? You can reach me on the Fediverse where I'm @lawik@fosstodon.org or by responding to this email to lars@underjord.io.

Thank you for reading. I appreciate it.