Compositional pipelines about guides

[pbrisbin]

To add some color for my own stance: I prefer C because I like to distinguish compositional pipelines (f . g . h) from application pipelines (f $ g $ h x). In other words, if an expression is pointful, it should be (simple) applications all the way, not an (unnecessary) applied composition, which is just a hair more complex to my brain. Totally subjective, but that's why.

[tfausak]

For some reason GitHub notifies me about new issues on this repo. And I can't help but comment on Haskell style questions 😆

I like style A because it's tied for the fewest number of operators and I find (.) much easier to visually parse than ($). However it can be really annoying to refactor that style because changing the number of functions involved can require changing a few operators; looking at things from that angle, style C seems like the best choice.

[cdparks]

I prefer C now, but I used to prefer E

E

f (g (h x))

Which probably means I can learn to prefer anything.

[cdparks]

For some reason GitHub notifies me about new issues on this repo

hiiii @tfausak 👋. I'm happy to see you chime in, but let us know if you want us to stop pinging you, should be easy to fix 😛

[tfausak]

Oh I'm sure I could find the setting to toggle, but I like these discussions. I've been trying to whittle away at our own style guide. Case in point: EdutainmentLIVE/haskell-style-guide#37

[5outh]

I prefer A (f . g $ h x)

• This uses the number of operators as C
• In general, I prefer function composition over application

I should also note that I don't mind if these styles are mixed. They're all acceptable to me (except D, but I think hlint will complain about that)

[jdreaver]

I use all of them depending on context and intent. I'm 👎 on having this in a style guide.

[pbrisbin]

I'm genuinely curious what intents/contexts would lead you to using either of these equivalent forms over others. Can you expand on that?

[jdreaver]

If I want something to feel "pipeline-y", like I'm doing a lot of complex operations in a row, I tend to use . a lot. If I just want to do a couple operations on something, like head $ take 2 $ [1, 2, 3, 4], I might just use $.

To be honest I don't have a nice, consistent rule for when I use one or the other because I don't think about it too much. I've seen all of these styles used and I don't think any one of them really has an advantage over the others since it is pretty clear what is going on.

[pbrisbin]

I don't think any one of them really has an advantage over the others

Just to be clear, no one is saying that's the case. It's just about consistency, and having a simple answer for people ramping up when they ask about how to write things.

[eborden]

I prefer A

[eborden]

Also a 👎 on a style guide for this.