4 ms·

It's notable that raku (previously known as perl6) lets you write things in either direction, if I remember correctly something like @source >>> map { ... }

by mst 4y ago

It's notable that raku (previously known as perl6) lets you write things in either direction, if I remember correctly something like @source >>> map { ... } >>> grep { ... } >>> my @sink; though note I'm typing from memory on my second coffee so I may have got that slightly wrong. Plus of course there's many languages with a |> operator so you can do g(x) |> f I also (the example is specialised for I/O but the implementation technique could trivially be borrowed for something that wasn't) implemented something sort of in this direction for perl once: http://p3rl.org/IO::Pipeline http://p3rl.org/IO::Pipeline I'm not convinced that left-to-right is -always- the best option and prefer having the choice of both, but I wouldn't be at all surprised if a survey of developers found that if they could only pick one they'd pick left-to-right, and while I'd find it hard to choose for myself alone I'd probably pick left-to-right on the basis that it'd likely make it easier to onboard people to any given codebase.

tmtvl 4y ago
Raku also has map and grep methods for collections, so it could be written as... my @sink = @source.map({...}).grep({...}).sort; Which also makes multithreading the operation easy: my @sink = @source.hyper.map({...}).grep({...}).sort.list; That said, I think the operator you're looking for is the feed operator, ==>: my @sink; @source ==> map {...} ==> grep {...} ==> sort ==> @sink; It also has the corresponding reverse, <==, because why not. The docs* mention that ==> may at some point do automatic parallelization, but I don't know what the status of that feature is. * https://docs.raku.org/routine/==%3E https://docs.raku.org/routine/==%3E
- mst 4y ago
  > That said, I think the operator you're looking for is the feed operator, ==> Yes, yes it was. Your clarifications, corrections and elaborations are much appreciated.

by mst 4y ago

tmtvl 4y ago

Raku also has map and grep methods for collections, so it could be written as... my @sink = @source.map({...}).grep({...}).sort; Which also makes multithreading the operation easy: my @sink = @source.hyper.map({...}).grep({...}).sort.list; That said, I think the operator you're looking for is the feed operator, ==>: my @sink; @source ==> map {...} ==> grep {...} ==> sort ==> @sink; It also has the corresponding reverse, <==, because why not. The docs* mention that ==> may at some point do automatic parallelization, but I don't know what the status of that feature is. * https://docs.raku.org/routine/==%3E https://docs.raku.org/routine/==%3E

mst 4y ago
> That said, I think the operator you're looking for is the feed operator, ==> Yes, yes it was. Your clarifications, corrections and elaborations are much appreciated.