I don’t know either. I guess Chris will tell us if he’d rather we did things a different way!
I started playing with it when I was working for a company that did “caching in the cloud” for speeding up web apps. It was basically a custom distributed multi-tenant caching system that looked like memcached to the users, but was running on a load of machines on AWS for better performance and redundancy. All the (very complicated) back-end code was in Go, and I thought it would be an interesting exercise to try rewriting some portion of it in Rust. It turned out to be just about the most unsuitable first Rust project imaginable. The whole thing is about highly optimised management of memory, and there are data structures with pointers pointing to pointers that point to more pointers that point back to the original pointers, and so on ad infinitum. That turns out to be hard to deal with in Rust, and would have required a lot of up-front design thinking to do things in a more principled way to make the Rust compiler agree that everything was okay. I failed fast there.
As for an embedded target, one will be an STM32 of some kind, mostly just because I have a Nucleo dev board lying around I can use to start with. That has an STM32F767 on it, which should have more than enough space and power for anything I want to do with it, and I can downsize from there once I’ve scoped out the software requirements for the project I’m doing. I’m doing another thing at the moment which is a tools comparison for the nRF52840. That’s slow work and slightly dull, so I’m keeping in mind the idea of doing some very basic Rust experiments on it at the end to keep me going!
Writing parsers in Rust sounds like a much more sensible project to start on than what I tried! I don’t know about these “half natural language” things. I had a little look at electro-grammar (which is nice) and Duckling, and Duckling at least mostly seems to be about spotting stereotyped phrases. I didn’t see anything in their examples that you couldn’t write a regular expression to match.