|
|
|
|
|
|
by wahern
2235 days ago
|
|
|
Unfortunately a substantial number of people seem to suffer from a type of concurrency Stockholm syndrome. They prefer explicit and awkward concurrency constructs, perhaps because until Go, mainstream languages like JavaScript, C#, and Python implemented hacks largely because of implementation constraints, so it's all they know. They say thinks like, "I prefer knowing exactly when I might block". Yet you never hear them bemoan the fact that the OS hides process scheduling and VM paging; they don't complain that they can't hook into this at all, let alone be forced to do it 100% of the time for even the simplest code. They never ask that their language force them to be explicit about how and when a simple function call will setup a new stack frame. All of these things were, decades ago, the equivalent of fibers (or more generically asymmetric stackful coroutines). People were likewise skeptical about those things, but today these innovations--function calls, preemptive scheduling, virtual memory--are ubiquitous and invisible and the notion of doing away with them completely or even as default semantics, for all basic programming, would be inconceivable. What small cost they impose upfront by being the default (and often only) construct is vastly outweighed by their power; they're undeniably the correct way to model the vast majority of computing problems at the level with which they're concerned. I agree, goroutines (or perhaps something a little more generic that doesn't conflate stack reification with scheduling) are absolutely the correct abstraction, and this will be proven in a few decades time. But until then we're stuck with people rationalizing the limitations of async constructs in their favorite languages; constructs primarily chosen because of quirky design constraints, principally C interop and C-based VM implementations that leak the semantics of the C ABI stack. |
|
|