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 fo

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.