You need two different compilers, and the ground is more fertile the more compilers you have. It would be more applicable to standards-based languages with a ro

You need two different compilers, and the ground is more fertile the more compilers you have. It would be more applicable to standards-based languages with a robust ecosystem of implementations like Common Lisp or SQL. Less so languages where there's mostly just the one compiler, like Scala or Rust. You can still use it to test different settings on a single compiler, though. Like optimization settings or target architecture.