5 ms·
It's interesting looking back on how this compared to the lessons learned by Rust in the lead up to 1.0. Issues in which Rust observed the same problems and ch
by pcwalton 10y ago
It's interesting looking back on how this compared to the lessons learned by Rust in the lead up to 1.0.
Issues in which Rust observed the same problems and changed course to fixed them:
- Compilation model: Rust originally wanted to do a static compilation model, like BitC, but it eventually went with a hybrid model that leans toward link-time optimization.
- Compiler-Abstracted Representations vs. Optimization: Again, Rust wanted to abstract representations from the compiler, but eventually it just went to a model in which the representations aren't hidden from the compiler.
- Insufficiency of the Type System: Rust also had by-ref parameters only, but as that was insufficient it eventually grew the lifetime system (as well as the subtyping needed to make it usable).
Issues in which Rust persevered anyway:
- Inheritance and Encapsulation: Rust went with interfaces and unified them with the typeclass system.
- Instance Coherence and Operator Overloading: Rust went with a Haskell-like approach to instance coherence. It is sometimes annoying, but it's rarely an insurmountable problem in practice, and Rust is a lot more flexible than most other non-C++, non-D languages here.
- scott_s 10y agoSome of this you covered in your comment in that thread (https://news.ycombinator.com/item?id=3750882 https://news.ycombinator.com/item?id=3750882). Is there anything in the past four years that you learned since that comment?