14:27
<Richard Gibson>
I can see how call-this on its own improves DCE (because it's easy for libraries to add exports of fields from classes that they're already exporting), and how a point-free pipe operator does so to a much lesser extent (by nudging the ecosystem towards compatible functions), but I'm having a real hard time with the suggestion that the chosen pipeline style will have any such effect, either on its own or even beyond call-this if they both make it. I like pipeline, but to me it seems independent of DCE. Can someone here help me connect the dots?
21:32
<Evan Winslow>
  1. (a few/some/many/most?) method users mainly value the left to right writing and reading order of method calls. 2. Hack pipeline provides left to right reading and writing order for plain functions. 3. Plain functions are more tree shakable than class methods.
22:39
<Richard Gibson>
but AFAICT, hack pipeline also provides the same benefits for method calls
22:40
<Richard Gibson>
i.e., it doesn't seem to incentivize plain functions
23:41
<Evan Winslow>
You mean because you can do something like expr |> ##.method()?
23:49
<Richard Gibson>
yes
23:54
<Evan Winslow>
Here's another way to think of it: Today methods score 5/5 for ergonomics/capabilities. Functions score 2/5 (these numbers are completely made up). With hack pipeline, maybe functions' score could increase to 3/5. So the relative incentive to use them would improve. On the margin, this will persuade some people to switch over from classes to functions.