Newest at the top
| 2026-02-23 21:53:23 +0100 | <EvanR> | oh wait |
| 2026-02-23 21:53:18 +0100 | <tomsmeding> | which I always find a funny phrasing, almost certainly chosen because it's the shortest way to express the intended set |
| 2026-02-23 21:53:11 +0100 | <EvanR> | given an odd non-1 prime... |
| 2026-02-23 21:52:49 +0100 | <tomsmeding> | there's plenty of math that assumes odd primes |
| 2026-02-23 21:52:33 +0100 | <EvanR> | so you don't have to preface everything you say about primes with "assuming no 1s" |
| 2026-02-23 21:51:44 +0100 | <lambdabot> | Infinity |
| 2026-02-23 21:51:42 +0100 | <humasect> | > (1 / 0) |
| 2026-02-23 21:51:32 +0100 | <tomsmeding> | "assuming these are all numbers" |
| 2026-02-23 21:51:31 +0100 | <EvanR> | another case of math beating us, by defining 1 to be not a prime |
| 2026-02-23 21:51:11 +0100 | <EvanR> | I have to preface everything I say about floats with "assuming no NaNs" |
| 2026-02-23 21:51:04 +0100 | <lambdabot> | NaN |
| 2026-02-23 21:51:02 +0100 | <tomsmeding> | > (1 / 0) + (-1 / 0) |
| 2026-02-23 21:50:32 +0100 | <tomsmeding> | NaN, however, breaks everything it touches; EvanR perhaps this influences what you observe |
| 2026-02-23 21:50:28 +0100 | <EvanR> | I approve |
| 2026-02-23 21:50:09 +0100 | <lambdabot> | True |
| 2026-02-23 21:50:07 +0100 | <tomsmeding> | > (1 / 0) * 2 > 3 |
| 2026-02-23 21:49:44 +0100 | <lambdabot> | 1.5707963267948966 |
| 2026-02-23 21:49:43 +0100 | <tomsmeding> | > atan (1 / 0) |
| 2026-02-23 21:49:35 +0100 | <haskellbridge> | <ijouw> atan (1/0) |
| 2026-02-23 21:49:23 +0100 | <tomsmeding> | infinity actually behaves kind of okay under certain floating point operations |
| 2026-02-23 21:49:23 +0100 | <haskellbridge> | <ijouw> I don't remember where it comes up and a value is more reasonable |
| 2026-02-23 21:46:47 +0100 | peutri | (~peutri@bobo.desast.re) peutri |
| 2026-02-23 21:46:12 +0100 | LUCKY_NOOB | (~LUCKY_NOO@user/LUCKY-NOOB:44374) LUCKY_NOOB |
| 2026-02-23 21:45:53 +0100 | LUCKY_NOOB | (~LUCKY_NOO@user/LUCKY-NOOB:44374) (Quit: leaving) |
| 2026-02-23 21:45:09 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 255 seconds) |
| 2026-02-23 21:45:08 +0100 | marinelli | (~weechat@gateway/tor-sasl/marinelli) marinelli |
| 2026-02-23 21:44:52 +0100 | peutri | (~peutri@bobo.desast.re) (Ping timeout: 246 seconds) |
| 2026-02-23 21:44:46 +0100 | marinelli | (~weechat@gateway/tor-sasl/marinelli) (Remote host closed the connection) |
| 2026-02-23 21:43:41 +0100 | <EvanR> | so annoying |
| 2026-02-23 21:43:31 +0100 | <EvanR> | also anybody get the feeling NaN comes up much more often than infinity for some reason |
| 2026-02-23 21:42:54 +0100 | machined1od | (~machinedg@d172-219-48-230.abhsia.telus.net) (Ping timeout: 255 seconds) |
| 2026-02-23 21:41:35 +0100 | <humasect> | div by zero feels more like inifnity |
| 2026-02-23 21:41:31 +0100 | <EvanR> | infinite == -infinite ok but infinite == zero might require some more coffee |
| 2026-02-23 21:41:17 +0100 | <haskellbridge> | <loonycyborg> in most cases infinity leads to exception |
| 2026-02-23 21:41:03 +0100 | <humasect> | ah well ... i would put infinity and zero in the same place. negative... |
| 2026-02-23 21:40:58 +0100 | <haskellbridge> | <loonycyborg> the answer is moo :P |
| 2026-02-23 21:40:20 +0100 | <haskellbridge> | <ijouw> infinity == -infinity ? |
| 2026-02-23 21:39:44 +0100 | <haskellbridge> | <loonycyborg> ye but it's still infinity |
| 2026-02-23 21:38:23 +0100 | <lambdabot> | (Infinity,-Infinity) |
| 2026-02-23 21:38:21 +0100 | <EvanR> | > (1 / tan 0.0, 1 / tan (-0.0)) |
| 2026-02-23 21:38:13 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-02-23 21:35:30 +0100 | <lambdabot> | Variable not in scope: cot :: t0 -> t1 -> t |
| 2026-02-23 21:35:28 +0100 | <EvanR> | > cot 0.0 (-1e-400) -- or even normal trig |
| 2026-02-23 21:34:46 +0100 | <EvanR> | and weeds upon weeds, the inverse trig functions care about the negative zero I think |
| 2026-02-23 21:33:37 +0100 | <EvanR> | it doesn't complain if everything ends up equal |
| 2026-02-23 21:33:16 +0100 | <EvanR> | that behavior wouldn't be problematic for x==y => f(x)==f(y) |
| 2026-02-23 21:32:12 +0100 | <haskellbridge> | <loonycyborg> like if we add negative or positive zero to something then we'll get exactly the same number |
| 2026-02-23 21:31:26 +0100 | <EvanR> | in the case of float it's kind of murky, what's the number itself, what's the representation (relevant to this law, and are reals even supposed to be like standard reals). Which I guess is food for thought when thinking about any other situation we are implementing |
| 2026-02-23 21:31:08 +0100 | <haskellbridge> | <loonycyborg> at least as needed in practice |
| 2026-02-23 21:30:46 +0100 | <haskellbridge> | <loonycyborg> positive and negative zero have different representations but represent same number |