welcome to null beach

day 1

1/26/2024

I've done some introductory market research into the current state of modernizing the embedded development experience, and it leaves so much wanting. There are several notable roadblocks to making changes:

• Vendor lock is a serious problem. Committing to a particular chip manufacturer and their proprietary toolchain has pros and cons, but when it comes to developer experience, closed-source toolchains are invariably worse than their open-source equivalents. Just let me use neovim like a sane person, why do I have to use your awful Eclipse clone and manage tons of IDE-specific artifacts just to build? Maybe I'm evangelizing a bit, but leveraging open source will grant your developers maximum flexibility to use the tools they are most familiar with, regardless of the platform.
• Embedded developers are notoriously resistant to change, scoffing at QoL improvements as passing fads. Some of them are. But this mentality limits more than it protects, in my opinion.
• The embedded development community is relatively small, fragmented, and "offline" - in comparison to our web/mobile/SaaS counterparts. The fact is that as the demand for pure software application development increases, pay will skew towards those jobs and skillsets, so less developers will enter the embedded workforce. Less young blood, less innovation, more comfortable greybeards writing C in vi until they retire. There's nothing wrong with that - but we can do better! It's a chicken-and-egg problem - embedded development requires patience, a special willingness to overlook ergonomics for the sake of getting things done. Many try it, and leave for a higher paycheck which requires far less initial effort.

There seems to be a good pretty opportunity in this space. There's a few companies that are trying their hand at modernization.

• Memfault

very interesting platform that emphasizes observability for IoT fleets. I gave it a cursory whirl - the UI is sleek, with lots of options. Administering a fleet of internet-connected devices through this UI looks like a pleasurable experience, and could deliver a ton of value to both technical and non-technical staff. They seem to have a pretty solid engineering staff and their blog is topnotch. Maybe I'll go work for them if this whole thing falls apart.

• Dojofive

• This is a small company based out of St. Paul, Minnesota, that purports to offer a similar modernization of development practices, but in the form of a full-blown consultancy.

• Open Source

There's a host of open-source frameworks like zephyr and platformio that are trying to move things in the right direction - but even these platforms are slow to be adopted.

My goal is to evaluate popular chip architectures with a series of increasingly complex Rust firmware projects until I settle on something worth pursuing to fabrication. Along the way, I plan to document my findings, and bring modern software development practices into the fold as much as possible - Docker to isolate dependencies and provide an easily distributable build environment, (my flavor of) test-driven development, CI/CD, source control, etc. If, along the way, I identify a gap in Rust support that requires open source contribution, I'd be thrilled to attempt to fill that gap as a method of gaining credibility in the community and contributing back to a language that genuinely think is the future of all systems computing

Even if I don't get rich, at least I learned something new.