Dan has blogged about this at length and given interviews on it. It's worth looking into. No language or program is ever perfect, so they are taking the approach that error recovery is a better approach than aiming for perfection. They originally were going with Wren, but it wasn't performant enough. C++ compiled to Web Assembly seems to be super performant and that's what this is all about. Also, there are ways to do type checking which make it pretty solid. Again, see Dan's posts about this stuff. Super interesting.
You are viewing a single comment's thread from:
First of all, since the web assembly will be the target language; I assume, we won't be tied to C++. We can use any language as long as it compiles to the web assembly. This is good news.
On the other hand, I agree there's a tradeoff and that's exactly my point. I think safety and correctness are more important for a smart contract and a language which is easier to reason about would be a better choice. Especially if that platform allows the contracts to be proved correct formally. Functional languages are not that slow with current machines and they lend themselves better for parallel execution.
I'll check Dan's posts. Thanks.