When comparing D vs Rust, the Slant community recommends Rust for most people. In the question“What is the best programming language to learn first?” Rust is ranked 21st while D is ranked 28th. The most important reason people chose Rust is:
Since Rust is statically typed, you can catch multiple errors during compile time. This is extremely helpful with debugging, especially compared with dynamically typed languages that may fail silently during runtime.
Specs
Ranked in these QuestionsQuestion Ranking
Pros
Pro Has an improved C subset
With few exceptions, D will either compile C code to work exactly as when compiled as C, or it won't compile - it won't quietly change the semantics of C code and give unexpected results. This means that D contains an improved C, as it fails compilation where type safety is missing in C.
This allows learning the same machine operations available in C and other low-level languages.
Pro Easy to read and understand code
Pro Doesn't force you to deal with memory management
When you're just starting out, dealing with manual memory management and its bugs is a huge pain! D is garbage collected by default, which removes a huge class of things that could go wrong. And later on, if you decide you don't want or need the GC, you can turn it off.
Pro Very fast compilation
D is usually up to 10 times faster than C++. Having a language that compiles this fast means that you are free to write highly optimized code because of the relatively low cost of experimentation.
Pro Unit testing built-in
D provides unittest blocks to insert code that verifies functions preform their expected behavior and document edge cases. Since these are enabled with a compiler switch, there is no need to teach new programmers how to install new libraries and import modules, instead training on test driven design can start from the very first function.
Pro Provides a powerful data structure on top of C's arrays called slices
D provides a structure that builds on C's arrays called slices. A slice is a segment of an array that tracks the pointer and the length of the segment.
Slices are extremely powerful because they combine the protection of knowing the length of the data with the garbage collector that manages the memory backing that data, thus avoiding most memory corruption issues.
Pro It's a state-of-art evolution of C
Pro Static with type inference
For a new user adding types can feel tedious, and takes focus off the meaning of the code, but they are also important for checking logic. D provides static types, and a good system to infer types, so types are checked when calling functions, but do not need to be specified everywhere, making it feel more dynamic.
Pro Provable purity and immutability
The compiler can check that functions don't have side effects, extremely important for functional programming in concurrent scenarios, and can check immutability.
Therefore, the compiler will prove that your programs are functionally pure and respect immutable data, if you want it to.
Pro Compile-time Function Execution
Pro Built-in Unicode support
Pro Industrial quality
Pro Asynchronous I/O that doesn’t get in your way
Because all types can be treated as objects, all files can call functions in the same manner -- even stdin
and stdout
. stdout.writeln();
stdin.readln();
file.writeln();
file.readln();
Pro Easy to integrate with C and C++
D practically has the same memory structure as C and C++; all D does it build a structure around that. The entire C standard library is accessible at no cost (syntactic or speed) and it's being worked on allowing the same for the C++ standard library.
Pro Designed for concurrency and parallelism
Supports first-class functionality for both concurrency and parallelism, offered as part of the standard library.
Pro Supports calling functions from types in an object-oriented manner.
if (exists(file)) {}
may be written as if (file.exists) {}.
writeln(file);
may be written as file.writeln();
isDivisibleBy(10, 2);
may be written as 10.isDivisibleBy(2);
writeln(isEven(add(5, 5)));
may be written as 5.add(5).isEven().writeln();
Pro Catch errors at compile-time
Since Rust is statically typed, you can catch multiple errors during compile time. This is extremely helpful with debugging, especially compared with dynamically typed languages that may fail silently during runtime.
Pro Compiles to machine code allowing for extra efficiency
Rust uses LLVM as a backend, among other things this allows Rust code to compile down to machine languages. This allows developers to write programs that run as efficiently as possible.
Pro Threads without data races
Unique ownership system guarantees a mutable data to be owned and mutated by only one thread at a time, so there's no data race, and this guarantee is checked at compile time statically. Which means easy multi-threading.
Of course, immutable data can be shared among multiple threads freely.
Pro Generics support
You don't have to write same array and dictionary classes hundreds and thousands times for strong type check by compiler.
Pro Built-in concurrency
Rust has built-in support for concurrency.
Pro Supports cross compilation for different architectures
Since Rust 1.8 you can install additional versions of the standard library for different targets using rustup/multirust
.
For example:
$ rustup target add x86_64-unknown-linux-musl
Which then allows for:
$ cargo build --target x86_64-unknown-linux-musl
Pro Makes developers write optimal code
Rust is a modern programming language written around systems. It was designed from the ground up this way. It's language design makes developers write optimal code almost all the time, meaning you don't have to fully know and understand the compiler's source code in order to optimize your program.
Furthermore, Rust does not copy from memory unnecessarily, to give an example: all types move by default and not copy. Even references to types do not copy by default. In other words, setting a reference to another reference destroys the original one unless it's stated otherwise.
Pro Support for macros
When you identify a part of your code which gets repeated often, which you cannot abstract using functions or classes, you can use Rust's built-in Macros.
Pro Official package manager and build tool
Cargo is the official package manager for Rust. It comes with the language and downloads dependencies, compiles packages, and makes and uploads distributable packages
Pro Easy to write understandable code
While not as verbose as Java, it still is much more verbose than languages like Go and Python. This means that the code is very explicit and easy to understand.
Pro Big community
The biggest community contributing to language.
Pro Functional programming
Very easy to create functional with some additional from structure application.
Pro Excellent syntax
Pro Safety
Pro No GC
Pro Low memory usage
Pro Static compiled & statically linked single file output
Pro Extremely fast code execution
Pro Zero-cost futures or Async IO
Cons
Con Poor adoption even after many years of existence
There's a widely accepted perception of D as a language that has been poorly adopted. Since adoption is driven by perception this becomes a fact. So managers and engineers start becoming nervous in adopting a language that has such a perception among the community and that has been so unsuccessful for so long.
Con Failed at becoming alternative to C or C++
Almost as confused and complicated as C++, but without the popularity and widespread corporate usage. Also failed at becoming a good cross-platform GUI application development language like Object Pascal. Many missed past opportunities, and now newer languages are better alternatives.
Con Lack of vision
D is community-driven and lacks the support of any large corporation. While this increases the amount of talent and engineering abilities of the people working on D, it also brings a severe lack of charisma, leadership and vision.
Con Garbage Collection
Memory is not managed directly.
Con All the downsides of garbage collection without any of its benefits
When D decided to implement garbage collection it instantly alienated a large community of developers (C and C++ programmers). For them, the official policy has been: "Don't want garbage collection? Use D with RAII or manual management style!".
While true, it's also absolutely pointless because there's little to none support for alternate memory management styles in the standard library, which means that a new user will have to start with a language that is stripped down of the core infrastructure.
On the other hand, for those people who want to use garbage collection, the implementation of it is lackluster.
Con Long compile times
Way longer to compile than other languages.
Con Low productivity
The compiler is strict, such as ownership and borrowing checkers.
Con Low readability
Harder to read and understand language.
Con Very ugly and verbose syntax
Compared to many languages. One tool needed to be productive is the language to be pleasing and effective, yet Rust friction is as high as its safety.
Con Steep learning curve
Pretty strict and hard to adapt for beginners.
Con Not as useful or productive as advertised.
If really out here wanting to use something besides C/C++, in 98% of use cases, would do just as well or be better off using something other than Rust. Will complete the project faster, would be easier, will be happier, and nearly as safe (use newer language with optional GC).
Con Rust not as safe as it pretends to be
Rust problems:
1) Massive undisclosed usage of unsafe.
2) Massive amounts of memory safety CVEs - see details.
3) Large bloated Rust binaries increase attack surface for exploits.
4) Many corporate types claiming Rust is safe, don't actually program or use it (have some quiet financial interest to push it).
Con Significant time required to understand the language fully
There's the infamous borrow checker for example.
Con Low-level programming language
This means that it encourages the programmer to be very careful in terms of how memory is allocated, etc.
Most applications can run without exceeding the capacity of the server, even with an inefficient dynamic scripting language.