Async/Await in C# and Javascript
  •  20 November 2018
  •  misc 

I recently changed from the NodeJS team to work on the .Net team (in the same company). Coming back to C# after a long time, there are a lot of new stuffs. Actually, I used to hate .Net (simply because I hate using Windows :LOL:). But thing has changed now. .Net Core can now run on non-Windows systems without any differences. Everything has become easier to develop .Net applications on Mac/Linux (using Jetbrains Rider like me or Visual Studio Community for Mac, which is a bad idea).

One interesting thing that I found in C# after a long time working in JS is the Async/Await operation, which simplifies asynchronous programming significantly. I heard that JS borrows the Async/Await idea from C#, so I decided to take a deeper look at the Async/Await operation in C# and compare it to the one in JS to see if there are any other things that C# is more successful at. There may be things that I was wrong about because I’m relatively new to C#.

Below is the comparison table between using Async/Await pattern in C# and JS. I also mentioned JS Generator because it can be applied pretty much in the same way as the one using Promise. Actually, it used to be an innovative way to solve asynchronous problems in JS before the birth of Async/Await. Many teams and products are still using it as the code base was developed many years ago. Today, Async/Await is the preferred way for handling asynchronous tasks in JS, leaving Generator back to its original purpose.

  C# JS (yield) JS (Async/Await)
Traditional Method TPL - Callbacks Callbacks Callbacks
Underlying Task Generator Promise
Invoke Immediately When calling yield Immediately
  The operation is executed immediately after calling the async method Calling a generator function returns a generator object only The operation is executed immediately after calling the async method
Postpone Execution Wrap in a Lambda expression No need Wrap in a function
Use Cases I/O & CPU-bound tasks I/O I/O
Multithread Use NO NO
    JS is single-threaded, Generator doesn’t help with CPU-bound tasks JS is single-threaded, Promise doesn’t help with CPU-bound tasks
Support Native, C# 5+ Native, but not all environments Native (newer JS versions)
    Requires 3rd-party libraries for async programming Use polyfill libraries in old JS environment
    co.js, js-csp  

Both the languages use Async/Await to solve a traditional problem with asynchronous programming, that is how to express the async tasks in a expressive synchronous style. The traditional methods use callbacks to handle the async tasks, which leads to a hard-to-maintain and hard-to-debug code base. The code is also spread into different contexts, which is error-prone and hard to reason about what is going on in the background. With Async/Await, the compiler (C#) or the run time platform (JS) take care of the hard work, leaving your application with a logical structure that resembles synchronous code.

C#, of course, is more powerful than JS in term of multi-threading support. It is not a surprise that C# supports both the I/O operations and the heavy CPU jobs while JS is only capable of I/O tasks. JS, by its nature, follows the single-threaded concurrency model, which relies on just one single main thread for doing all the work. JS is not suitable for running CPU intensive applications so Generator and Promise only provide a better way to organise the code. They don’t contribute anything to the performance.

There are still some other C# features that I haven’t taken a deeper look at and waiting for me to explore. It’s quite an interesting environment, from the platform, the language to the support and the tools (although I’m a bit allergic to its Capitalized style). There will be more posts like this if I have enough time :D

Update: as the time of writing this post, NodeJS has started to experiment Worker Thread, a module that provides a way to create multiple environments running on independent threads. Many browsers also come with Service Worker APIs, which have the similar abilities as Worker Thread.