Newest at the top
2025-04-28 01:47:37 +0200 | <EvanR> | or why you would want to |
2025-04-28 01:47:33 +0200 | <EvanR> | I don't see how you would even erase a type that is generated at runtime |
2025-04-28 01:47:01 +0200 | xff0x | (~xff0x@2409:251:9040:2c00:8240:4eb8:4326:3de4) (Ping timeout: 248 seconds) |
2025-04-28 01:46:41 +0200 | <haskellbridge> | <loonycyborg> like need to explicitly figure out what is supported and what isn't |
2025-04-28 01:46:28 +0200 | <mange> | Idris 2 lets you be precise about which types you expect to be erased from the program, using multiplicities: https://idris2.readthedocs.io/en/latest/tutorial/multiplicities.html#erasure |
2025-04-28 01:46:26 +0200 | <haskellbridge> | <loonycyborg> if you explicitly don't support it then it's unlikely that it's a localized case |
2025-04-28 01:46:18 +0200 | <EvanR> | you can remove them like a compiler is expected to remove terms that aren't doing anything |
2025-04-28 01:46:03 +0200 | <haskellbridge> | <loonycyborg> But still there is the case of reading of arbitrary type from stdin :P |
2025-04-28 01:44:54 +0200 | <EvanR> | once you get to a phase where the types aren't doing anything, then keeping them around can only waste resources |
2025-04-28 01:43:48 +0200 | <haskellbridge> | <loonycyborg> so to get proper dependent types in a compiled language would have to drop type erasure, perhaps replacing it by something else |
2025-04-28 01:43:19 +0200 | <EvanR> | because e.g. constant folding at compile time etc |
2025-04-28 01:43:01 +0200 | <EvanR> | well that's not true is it |
2025-04-28 01:42:33 +0200 | <haskellbridge> | <loonycyborg> to me type erasure seems like technical decision asserting that "compile type == type level; runtime == term level" |
2025-04-28 01:41:55 +0200 | <EvanR> | isn't a type level thing |
2025-04-28 01:41:46 +0200 | <EvanR> | accept keyboard input and append to an ongoing list of Types that you can then ask for a listing of, because they have nice Show instances |
2025-04-28 01:40:30 +0200 | <EvanR> | well then that's not purely at the type level definitely |
2025-04-28 01:39:50 +0200 | <haskellbridge> | <loonycyborg> I'm not so sure because of types that could be specified entirely at runtime |
2025-04-28 01:39:32 +0200 | <EvanR> | (but I heard java erases it's types also) |
2025-04-28 01:39:24 +0200 | mange | (~user@user/mange) mange |
2025-04-28 01:39:22 +0200 | <EvanR> | because this isn't java |
2025-04-28 01:38:54 +0200 | <EvanR> | if it's purely at the type level then it wouldn't need to exist at runtime |
2025-04-28 01:38:33 +0200 | <haskellbridge> | <loonycyborg> which probably will have different implementations at runtime and compile time |
2025-04-28 01:38:16 +0200 | <haskellbridge> | <loonycyborg> still it requires operating on type level things that can't be erased |
2025-04-28 01:37:20 +0200 | <EvanR> | erasing types or proofs shouldn't change the "observable" behavior |
2025-04-28 01:36:36 +0200 | fantom | (~fantom@2.219.56.221) (Ping timeout: 244 seconds) |
2025-04-28 01:36:08 +0200 | <EvanR> | typing with pinkies is hard |
2025-04-28 01:36:02 +0200 | <monochrom> | Maybe the programmer doesn't use that encoding, but the compiler does the translation. |
2025-04-28 01:35:56 +0200 | <EvanR> | if that's in the machine code then it only bothers somebody trying to disassemnle it |
2025-04-28 01:35:01 +0200 | <haskellbridge> | <loonycyborg> Even if you technically make this work not sure it would be a good language design. Like most types that aren't passed around are erased but some have to remain which kinda feels inconsistent. |
2025-04-28 01:34:21 +0200 | <monochrom> | :) |
2025-04-28 01:34:14 +0200 | <EvanR> | ok yes |
2025-04-28 01:32:36 +0200 | <monochrom> | Err, ∀r. (Πx:A. B(x) -> r) -> r |
2025-04-28 01:32:13 +0200 | <EvanR> | ... not sure |
2025-04-28 01:29:14 +0200 | <monochrom> | Do you accept: Encode Σx:A. B(x) as ∀r. Πx:A. B(x) -> r so it is just pi types all over again? :) |
2025-04-28 01:25:24 +0200 | sprotte24 | (~sprotte24@p200300d16f174f00e11b2faf6af92897.dip0.t-ipconnect.de) (Quit: Leaving) |
2025-04-28 01:24:43 +0200 | acidjnk_new | (~acidjnk@p200300d6e71c4f29a1deb7f42d1df083.dip0.t-ipconnect.de) |
2025-04-28 01:21:18 +0200 | <EvanR> | sigma type is a bit more mysterious to me, how do you know someone won't want to use the 2nd component |
2025-04-28 01:19:41 +0200 | <EvanR> | just discard the argument |
2025-04-28 01:19:29 +0200 | <EvanR> | any call site would need to be adjusted to stop trying to call it like a function |
2025-04-28 01:18:01 +0200 | peterbecich | (~Thunderbi@syn-047-229-123-186.res.spectrum.com) peterbecich |
2025-04-28 01:16:45 +0200 | j0lol | (~j0lol@132.145.17.236) j0lol |
2025-04-28 01:16:37 +0200 | <EvanR> | if the code doesn't use the argument, be it a type or a value, then the body is effectively a constant |
2025-04-28 01:16:04 +0200 | Typedfern | (~Typedfern@242.red-83-37-36.dynamicip.rima-tde.net) (Ping timeout: 260 seconds) |
2025-04-28 01:12:49 +0200 | typedfern_ | (~Typedfern@242.red-83-37-36.dynamicip.rima-tde.net) typedfern |
2025-04-28 01:10:25 +0200 | <haskellbridge> | <loonycyborg> I still have no idea how can you have both pi types and type erasure in the same language. |
2025-04-28 01:03:14 +0200 | koz | (~koz@121.99.240.58) (Ping timeout: 244 seconds) |
2025-04-28 01:02:21 +0200 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn |
2025-04-28 00:59:06 +0200 | peterbecich | (~Thunderbi@syn-047-229-123-186.res.spectrum.com) (Ping timeout: 244 seconds) |
2025-04-28 00:50:58 +0200 | peterbecich | (~Thunderbi@syn-047-229-123-186.res.spectrum.com) peterbecich |
2025-04-28 00:49:49 +0200 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |