On Friday, 26 June 2015 at 07:00:22 UTC, thedeemon wrote:
On Friday, 26 June 2015 at 04:43:08 UTC, Rikki Cattermole wrote:
Looks like I will give a talk about D to our local Functional
Programming User Group in August.
irk be careful when showing off ranges. Get some damn good
background of what D is first. It took me well over a year
before I started to get into them.
For FP people working with combinators on lazy lists or
sequences is a well known thing, so they should easily
recognize it.
People in OCaml, F# or Elm write code like this every day:
Char.toCode c |> spt |> List.reverse |> List.map toString |>
String.join " "
(this one is in Elm)
And in Haskell it looks similar but usually in backwards order
and separated by "." and $. For haskellers, comparing D ranges
with "classy prelude" shall be appropriate, I guess.
I think one should also talk about how D ranges when passed
through several processing steps can keep important information
in their types so that xs.retro.take(100).map!f.retro[33] works
in O(1). This is what most other languages don't do.
I absolutely _hated_ laziness in Haskell. It made it impossible
to reason about the performance of anything. I find it the
complete opposite in D because so much happens at compiletime,
I've seen ldc and gdc turn rather long range chains into assembly
you'd expect from nested for loops.
