Greg Colvin
Feb 4 · 1 min read

Excellent, thanks. I agree that the EVM would not have met your requirements, and “isomorphic blockchain logic” is a very exciting idea. Amd I’m very glad to see some of the security issues with compilers and JITs explained — security is job #1 on a public blockchain. I do think that a secure, near-linear-time (n log n) Wasm compiler, having foregone complex optimizations, should not be that difficult to write. Wasm was designed with low-level opcodes and validated control flow so that much of the hard work of an optimizing compiler can be performed on the Wasm code itself, with the Wasm compiler making one pass over the result and doing simple code generation.