Newest at the top
| 2025-01-28 23:09:46 +0100 | <tomsmeding> | ah, that's wrong. Set `data T = T A`. |
| 2025-01-28 23:09:25 +0100 | <tomsmeding> | if the function goes A -> B, set T = A. |
| 2025-01-28 23:08:58 +0100 | <monochrom> | Related unhelpful theorem: For every computable function, there exists an intermediate data structure T such that the function is composed of ana'ing input into T and then cata'ing T into output. |
| 2025-01-28 23:08:43 +0100 | img | (~img@user/img) (Ping timeout: 245 seconds) |
| 2025-01-28 23:07:24 +0100 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn |
| 2025-01-28 23:07:19 +0100 | <tomsmeding> | agreed |
| 2025-01-28 23:07:19 +0100 | confused_rust_en | (~confused_@2603:7081:1600:6425:187f:fcc6:e582:3131) |
| 2025-01-28 23:07:02 +0100 | <int-e> | The serious part though is that when figuring out algorithms on simple data structures... "what recursion scheme is this" doesn't seem like a helpful question. |
| 2025-01-28 23:06:51 +0100 | <haskellbridge> | <alexfmpe> Might need some unsafe functions in there for traverse-ish if you only require Foldable |
| 2025-01-28 23:06:18 +0100 | <int-e> | It does get interesting when you compose these things... one early showcase is build/foldr fusion. |
| 2025-01-28 23:06:09 +0100 | img_ | (~img@user/img) img |
| 2025-01-28 23:05:49 +0100 | <haskellbridge> | <alexfmpe> You can write all the other ones via it |
| 2025-01-28 23:05:35 +0100 | <haskellbridge> | <alexfmpe> Fold *is* recursion |
| 2025-01-28 23:05:33 +0100 | <tomsmeding> | I guess foldMap is a little less powerful than a full catamorphism would be |
| 2025-01-28 23:05:06 +0100 | <monochrom> | I consider foldMap to be under monoidology not recursion schemes. >:) |
| 2025-01-28 23:04:55 +0100 | <mauke> | map is quite useful |
| 2025-01-28 23:04:36 +0100 | <tomsmeding> | int-e: foldMap is quite useful. |
| 2025-01-28 23:03:13 +0100 | machinedgod | (~machinedg@d108-173-18-100.abhsia.telus.net) machinedgod |
| 2025-01-28 23:02:07 +0100 | <int-e> | Recursion schemes are silly; explicit recursion is underrated. Change my mind. |
| 2025-01-28 23:00:50 +0100 | <monochrom> | Ah, I need to respect that. But only out of a mathematical/logical reason: In type theory with user-definable algebraic types (e.g. Agda, Lean), the auto-gen'ed induction principle is a para. |
| 2025-01-28 22:59:43 +0100 | <EvanR_> | the effect of code on humans and vice versa |
| 2025-01-28 22:59:01 +0100 | <EvanR_> | monochrom, stuff like recursion schemes might fall under programming geography |
| 2025-01-28 22:57:31 +0100 | sawilagar | (~sawilagar@user/sawilagar) (Ping timeout: 252 seconds) |
| 2025-01-28 22:57:02 +0100 | Midjak | (~MarciZ@82.66.147.146) (Quit: This computer has gone to sleep) |
| 2025-01-28 22:56:39 +0100 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 2025-01-28 22:55:44 +0100 | tnt2 | tnt1 |
| 2025-01-28 22:55:44 +0100 | tnt1 | (~Thunderbi@user/tnt1) (Ping timeout: 265 seconds) |
| 2025-01-28 22:55:06 +0100 | <c_wraith> | (at least in a non-strict language) |
| 2025-01-28 22:54:58 +0100 | Googulator | (~Googulato@2a01-036d-0106-1666-e945-fd21-b920-9aa7.pool6.digikabel.hu) (Client Quit) |
| 2025-01-28 22:54:48 +0100 | <c_wraith> | monochrom: paramorphisms can be an efficiency win, even if they have the same computational power as catamorphisms. |
| 2025-01-28 22:53:55 +0100 | <tomsmeding> | :) |
| 2025-01-28 22:53:38 +0100 | <haskellbridge> | <thirdofmay18081814goya> thanks a lot for the example! |
| 2025-01-28 22:53:33 +0100 | <tomsmeding> | to do better in terms of space usage, you'll have to augment the folder |
| 2025-01-28 22:53:33 +0100 | <haskellbridge> | <thirdofmay18081814goya> hm I see |
| 2025-01-28 22:53:16 +0100 | <tomsmeding> | goya: note that my code has to reconstruct the tree in the fold, even though there is already one in memory: the one we're folding over. So this returns new copies of each subtree instead of shared subtrees |
| 2025-01-28 22:53:10 +0100 | <int-e> | strictly awful? |
| 2025-01-28 22:52:42 +0100 | tnt2 | (~Thunderbi@user/tnt1) tnt1 |
| 2025-01-28 22:52:23 +0100 | <monochrom> | Oh! To avoid lazy evaluation, just lift to liftA2 (&&) for a suitable Applicative instance. >:) |
| 2025-01-28 22:52:08 +0100 | <tomsmeding> | use Maybe instead of [] to get only the first / last / ... match |
| 2025-01-28 22:52:03 +0100 | <haskellbridge> | <thirdofmay18081814goya> will be studying that example |
| 2025-01-28 22:52:00 +0100 | <haskellbridge> | <thirdofmay18081814goya> oh neat |
| 2025-01-28 22:51:38 +0100 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn |
| 2025-01-28 22:51:32 +0100 | <monochrom> | Stops early. |
| 2025-01-28 22:51:31 +0100 | <c_wraith> | Yeah, you don't need to work for that. You need to work to avoid it. |
| 2025-01-28 22:51:27 +0100 | <lambdabot> | False |
| 2025-01-28 22:51:25 +0100 | <monochrom> | > foldr (&&) undefined (False : undefined) |
| 2025-01-28 22:51:15 +0100 | <tomsmeding> | what monochrom says |
| 2025-01-28 22:51:08 +0100 | <monochrom> | Lazy evaluation easily stops early when something is found. |
| 2025-01-28 22:51:05 +0100 | <tomsmeding> | goya: https://play.haskell.org/saved/KMCLVKww |
| 2025-01-28 22:50:47 +0100 | Googulator | (~Googulato@2a01-036d-0106-1666-e945-fd21-b920-9aa7.pool6.digikabel.hu) |