Re:Rust's Complexity Problem – A Warning
by chiefsucker on Wednesday March 18, 2020 @04:34AM
(#22614696)
> It is obvious the author has little to no experience in developing production software in Rust.
Statements like these don’t necessarily speak for the user-friendliness of Rust or Rust itself. Safety? OK, cool. But for what cost?
Re:Rust's Complexity Problem – A Warning
by jacobr1 on Wednesday March 18, 2020 @11:46AM
(#22617914)
I think the cost is roughly: "we force you to reason about, and account for safety in the code." Then given that this is now explicitly forced upon the developer, are Rust's constructs overly complex, or reasonably once someone gets over the learning curve? The alternative is using a managed runtime, which probably should be the default for most programmers. But for those that are doing systems programming, we absolutely need something "like Rust." The memory access patterns need to be explicit.
Re:Rust's Complexity Problem – A Warning
by unlinked_dll on Wednesday March 18, 2020 @12:23PM
(#22618326)
Personally I think the type system (via traits and genetics) makes it really easy to build abstractions with low costs and friendly APIs, which makes composing good software that runs faster much easier.
There seems to be a class of developer that's been brainwashed into thinking generic programming is bad. It's pretty damn great when your type system is well designed.
This is a problem for newcomers because they have to adapt their way of thinking into Rust way of thinking. Once one develops a set of patterns and principles this is not an issue.
Companies who have large codebases in Rust tend to have praise for it enabling solutions that are much easier to reason with and work on then if they were developed in C++.
It is obvious the author has little to no experience in developing production software in Rust.