> but this is the only right way of evolving a programming language when incompatible changes must be made: by providing tools that guarantee [it] Yep. I'm no

> but this is the only right way of evolving a programming language when incompatible changes must be made: by providing tools that guarantee [it] Yep. I'm not a particularly big fan of Google languages, but focusing on bridging legacy code is brilliant. Would have been interesting if Oracle had done the same with Java shortly after acquiring it. Java is so hemmed in by legacy compatibility it basically can't make any significant language or VM changes without making harsh compromises. Of course, Oracle would likely have supported legacy Java indefinitely, and made a fortune supporting/licensing it.