#Rust2019 blog post: The Last Thing Rust Needs


#41

It’s interesting that while let itself could be seen as a case where added sugar codifies a special case and thus disincentivises implementing a general solution to the problem (pattern matching expressions, see 1, 2, 3).


#42

The fundamental problem might be, that you don’t get that kind of feedback until more people have used the feature, which might be only the case if it’s on stable.

Yes, it’s important to not add features willy-nilly - which certainly isn’t the case for Rust - but being able to deprecate features and in the end remove them after some time, might be long-term a lot more important.

So I’m very happy to see the work on Rust editions and tools like ‘cargo fix’ which go into this direction.

You really can’t foresee the evolution of a programming language. You want to be able to step forward without to much fear, being able to experiment and go into new directions. But also to be able to step backward again, if something didn’t work out, without introducing to much harm and work.

It’s really hard to get this right.