The most important feature that attracts me to Rust is its focus on allowing the language to grow.
If Rust was going to remain static as it currently exists, I would probably have avoided it and chosen a different language to spend my time with, such as OCaml, Haskell, or Racket.
My two main concerns regarding code, and what influences the things I study, are:
- How fast will I be able to go from idea to implementation?
- How maintainable will the code be? Will I be able to make changes easily?
However, Rust brings me the following benefits:
- Control of the hardware. I learned C as a student ages ago by reading K&R, and I know that for me to write production-quality C would require lots of ramp up time.
- Speed. Even if my naive code isn't as fast as it could be, I feel reasonably sure that I can make whatever speed improvements I need without making the code unreasonably eldritch.
These sets of benefits don't intersect. So, why am I interested in Rust?
I see Rust as a starting point, a point from which other, higher level abstractions can be added. I have heard numerous times that features like garbage collection will be added after the fact, and that this is can be implemented as user-land libraries. Indeed, much of what would normally be in the standard library is being put into crates, allowing libraries to grow independent of Rust itself.
Racket, in contrast, is an amazing system for exploring and implementing new ideas in syntax, semantics, and abstraction. However, its runtime and base system mean that it is suitable for a relatively select domain of problems. These domains have large, established language communities (Ruby, Python) that are losing mind share because of these limitations, not because of a lack of expressiveness in those languages.
Because of these considerations, Rust is a good base language that can grow into the future. Its focus on not making trade-offs that preclude certain programming domains means that we can use it to starting building better systems, be they applications, operating systems, or new programming languages with a restricted domain of applicability.