Aren't the Ocaml and ReasonML syntaxes so similar that you can effectively convert both directions and still end up with clean idiomatic code? That has seemed to be the case when I've played around with the ReasonML, and if it's the case, then which one you choose to write is very much arbitrary. Heck, you could even write one syntax locally and then automatically convert to the repo's preferred syntax when you commit your code.