time-travelling search can be done with recursion. Otherwise, i agree that they don't compose well with each other in some languages. However, it is harder to implement efficiently as far as i understand; time will tell if they are widely understood to be implemented in future languages.
Many languages have exceptions, but with shift/reset, you can implement your own exception handling behavior trivially, including things like resumable exceptions.
Some languages have async/await, but almost all of those implementations are state-machine based, which means that they don't compose well when used with more complex language features like generics and higher order functions.
But more than that, in continuation-based asynchronous programming, "async" functions are just normal functions that execute in an async prompt. Here's an example of a multi-user asynchronous TCP echo server in my toy language with tagged delimited continuations:
Where all of those asyncrhonous constructs (including `async` itself) are really short wrappers around delimited continuation machinery.You'll notice that the nested calls to `async` indicate that there are two asyncrhonous contexts
1. Connection acceptance happen on their own "thread" of control
2. Each connection gets their own context in which everything is internally asyncrhonous