Newest at the top
| 2026-02-20 18:49:40 +0100 | n0w0n | (~john@user/n0w0n) (Ping timeout: 268 seconds) |
| 2026-02-20 18:49:27 +0100 | housemate | (~housemate@202.7.248.67) housemate |
| 2026-02-20 18:49:05 +0100 | housemate_ | (~housemate@202.7.248.67) (Quit: https://ineedsomeacidtocalmmedown.space/) |
| 2026-02-20 18:48:08 +0100 | ChanServ | +v yahb2 |
| 2026-02-20 18:48:08 +0100 | yahb2 | (~yahb2@user/tomsmeding/bot/yahb2) yahb2 |
| 2026-02-20 18:47:46 +0100 | yahb2 | (~yahb2@user/tomsmeding/bot/yahb2) (Server closed connection) |
| 2026-02-20 18:47:18 +0100 | <haskellbridge> | <ijouw> Which acts as multiway Either |
| 2026-02-20 18:46:50 +0100 | <haskellbridge> | <ijouw> No i mean introducing a type SumT :: Set Type |
| 2026-02-20 18:46:11 +0100 | <Leary> | Well, not quite. A `Dynamic` limited to `Ord`. |
| 2026-02-20 18:45:58 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 268 seconds) |
| 2026-02-20 18:45:39 +0100 | <Leary> | ijouw: Re your original question, I'm not entirely sure what you're asking, but `Dynamic` could (and should) have `Ord`. |
| 2026-02-20 18:45:27 +0100 | Vizious | (~bes@user/Vizious) (Quit: WeeChat 4.8.1) |
| 2026-02-20 18:45:11 +0100 | gmg | (~user@user/gehmehgeh) (Quit: Leaving) |
| 2026-02-20 18:44:09 +0100 | <Leary> | `(* (a :+ 0))` and `(+ (a :+ b))` are both monotone wrt the lexicographic ordering GHC would derive. |
| 2026-02-20 18:42:16 +0100 | AlexZenon | (~alzenon@85.174.181.228) (Ping timeout: 268 seconds) |
| 2026-02-20 18:41:39 +0100 | tzh | (~tzh@c-76-115-131-146.hsd1.or.comcast.net) |
| 2026-02-20 18:40:58 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-02-20 18:39:59 +0100 | <haskellbridge> | <ijouw> *circle |
| 2026-02-20 18:39:39 +0100 | <haskellbridge> | <ijouw> Likely using angles on the unit sphere and magnitude. |
| 2026-02-20 18:38:30 +0100 | <haskellbridge> | <ijouw> But I am too lazy to formalize it. |
| 2026-02-20 18:37:59 +0100 | <haskellbridge> | <ijouw> There is one order where (*2) does not affect order, but it will fail if you add something. |
| 2026-02-20 18:37:15 +0100 | <c_wraith> | which is documented to produce broken output if the transformation isn't monotonic with respect to the ordering |
| 2026-02-20 18:36:55 +0100 | AlexZenon | (~alzenon@85.174.181.228) |
| 2026-02-20 18:36:49 +0100 | <c_wraith> | fwiw, Data.Set has mapMonotonic |
| 2026-02-20 18:33:44 +0100 | lisbeths | (uid135845@id-135845.lymington.irccloud.com) lisbeths |
| 2026-02-20 18:33:06 +0100 | lisbeths | (uid135845@id-135845.lymington.irccloud.com) (Server closed connection) |
| 2026-02-20 18:31:57 +0100 | <haskellbridge> | <ijouw> But yes, doing something like Set.map (*2) over complex numbers will rebuild the tree (for any total Ord and random values). |
| 2026-02-20 18:30:19 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 264 seconds) |
| 2026-02-20 18:29:00 +0100 | <haskellbridge> | <ijouw> The empty one |
| 2026-02-20 18:27:53 +0100 | <c_wraith> | I don't know what data structure *does* give you efficient set-membership testing and also supports arbitrary transformations of all values in the structure without rebuilding it |
| 2026-02-20 18:27:39 +0100 | <haskellbridge> | <ijouw> Well doing fmap conjugate does not work either way |
| 2026-02-20 18:26:41 +0100 | m_a_r_k | (~m_a_r_k@archlinux/support/mark) m_a_r_k |
| 2026-02-20 18:26:35 +0100 | <n0w0n> | Ig ya'd lose most of the benefits of doing complex ring operations to all members of a set without rebuilding a red black tree |
| 2026-02-20 18:26:26 +0100 | m_a_r_k | (~m_a_r_k@archlinux/support/mark) (Server closed connection) |
| 2026-02-20 18:24:55 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-02-20 18:21:46 +0100 | <n0w0n> | An incongruent ordering, ya say? |
| 2026-02-20 18:19:43 +0100 | kuribas | (~user@ip-188-118-57-242.reverse.destiny.be) (Remote host closed the connection) |
| 2026-02-20 18:19:06 +0100 | <dutchie> | just costs a newtype wrapper I guess |
| 2026-02-20 18:18:31 +0100 | <dutchie> | which doesn't matter if you just want to shove them into a Set |
| 2026-02-20 18:18:12 +0100 | <dutchie> | you can order the complex numbers (eg. lexicographically on their real and imaginary components), it just won't be compatible with the ring structure |
| 2026-02-20 18:16:23 +0100 | Enrico63 | (~Enrico63@host-82-56-209-65.retail.telecomitalia.it) (Quit: Client closed) |
| 2026-02-20 18:15:18 +0100 | skinkitten | (~skinkitte@user/skinkitten) (Quit: Client closed) |
| 2026-02-20 18:14:43 +0100 | <n0w0n> | Damn you Chris Okasaki |
| 2026-02-20 18:14:29 +0100 | <n0w0n> | & Sets are secretly Red Black Trees with an ordering |
| 2026-02-20 18:13:44 +0100 | <n0w0n> | Since complex numbers are unordered |
| 2026-02-20 18:13:39 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds) |
| 2026-02-20 18:13:35 +0100 | <n0w0n> | Unrelated, but that just reminded me that ya can't encode a set of complex number in Haskell using the Set type in containers |
| 2026-02-20 18:11:44 +0100 | <haskellbridge> | <ijouw> What if we made an arbitrary Ord Type and used it and Set Type to define dynamic sum types (as seen in e.g. typescript)? |
| 2026-02-20 18:10:58 +0100 | Googulator | (~Googulato@193-226-241-165.pool.digikabel.hu) |
| 2026-02-20 18:10:43 +0100 | Googulator | (~Googulato@2a01-036d-0106-499d-1c5a-0e4b-5795-d7e3.pool6.digikabel.hu) (Quit: Client closed) |