Another thing I highlighted in Rust 2018: The year we learn Rust 1.0 is the need to develop software practices at scale. I have a positive view on our progress there.
Crucial tooling like
cargo generate are in a good shape, a great example for this are the project setups in the embedded book. It allows us to pre-ship reasonable application structures. I’m a huge fan of it and hope for a wider adoption. In my book,
cargo generate should replace
cargo new --bin in most cases in the future.
But also, patterns developed, like the ownership patterns around drivers, as presented in James Munns talk at RustConf. But beyond embedded, discussions around high-level details have emerged. One great example here are Nikos office hours. Construction and composition of large systems is hard to explain in the void. They are often context sensitive. Taking a conversational approach by answering community questions out of our own experience seems like a good plan here.
With the upcoming stabilisation of
async/await, I hope that we will see a lot of patterns emerge here more clearly as well.