RISC-V Software Ecosystem

There will be a need to support legacy software for RISC-V, but since RISC-V is the real open and royalty free standard for ISAs, the real future is assembly for many software components.

But some of the mistakes from the legacy should be avoided: if the assembly code is dependent on expensive code generators or abuse is done to a macro pre-processor, this ain't that less worse than a compiler and bloated SDKs:

Then assembly code, without any code generator (or written in assembly themselves) and with a very conservative and minimal use of a very simple macro preprocessor would be the way to go.

Think Linaro, but for RISC-V. Each sponsor has to assign up to 2 full-time software engineers to the project who then work on RISC-V projects exclusively.

It's like good old ARM bringup, except this time most of the work is already done and we know how to do it.

And, this time, there's more good will than ever; We're not helping popularizing some singular company's ISA: We're helping the open ISA take off.

A lisp machine with an editor and seL4, Plan 9 goodies outperforming nvidia's ai devkit and looking just as good and scalable would be awesome to have.