On Sat, Jun 29, 2019 at 9:30 PM Henrik Johansson <dahankz...@gmail.com> wrote:
> > I have a feeling that there is a quite large "silent majority" that pretty > much agrees with me. > > Go is a language with the following properties, among other things: * Massive concurrency through a high amount of gorutines * Identity of channels, but not goroutines (goroutines don't have a Pid by construction, channels do have identity) * Shared memory space among the goroutines Such a language has a subtle property as an implication: cleanup of a failure must happen in the goroutine which made it. That is, error handling must be localized. As a result, error handling is mostly a game of treating it as data-flow in the program, where errors are values. There is a common case in such data flow. If an error occurs, we want to just return that error. This is where the try-construct will definitely improve the ergonomics of the programming. Go is a fairly small language. Certainly smaller than e.g., Javascript or Typescript. So I don't think we are adding that much cognitive load on people with a specialized construction. Of course YMMV... Haskell or OCaml programmers would just define a monad and call it a day, heh. -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/CAGrdgiVU42ep%2BzJFUHc8FiZ_yQWrw2YOTCJEjMbLkyiFeC_C%3DA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
