3 ms·
"Now that's interesting." I knew you'd like it. It's why I tried to message you in another comment about it. " Is it available for small embedded machines in
by nickpsecurity 11y ago
"Now that's interesting."
I knew you'd like it. It's why I tried to message you in another comment about it.
" Is it available for small embedded machines in the Arduino and low-end ARM class? It would be a good choice for when you just need to run some sensors and actuators, and need reliability."
Good point. I just found it thanks to another HN commenter. I don't know much about the state except to say they're jump starting it with another Modula-2-to-C compiler. I have the source for the original mtc compiler if anyone needs it. Meanwhile, the new one needs skilled contributors.
The language spec itself is in alpha and can use input on its features. They're adding stuff that might help. I'd say now is the time to suggest things that offer safety or maintenance benefits without bloating the language. Anyone you or anyone else can think of. I for one intend to go through Ada and Eiffel looking for every trick they use for safety for internal and integrations. Then, see what can be applied to Modula-2 without ridiculous effort. Design by contract (or a knockoff), existential types, and SCOOP concurrency come to mind immediately. Plus, maybe a formal semantics or a subset to aid static analysis tools or formal verification.
Macro's, too. I want safe (or safe-ish) macros. My old BASIC/LISP-based 4GL had macros for DSL's. They were awesome. So good for portability, security, performance hacks... many things with less overhead or confusion. Let's let people extend the language without punishing them by forcing knowledge of compiler writing.
Whats your thoughts on jumpstarting with a M2-to-C compiler, borrowing from Ada/Eiffel in simpler form, and adding macros?
"Go is for larger systems, in the large web server class. It does too much implicit allocation and garbage collection for small embedded work."
Oh I agree. I was just pointing out they created an Oberon-style language with industrial enhancements to solve problems of Java, etc. It worked with quite a bit of uptake. Like Oberon, the programs are pretty simple, safer, compile fast, and run fast. Old benefits carried over into that application of Wirth style. So, applying it to a system-level language will probably re-create the prior benefits in that area.