| --- Log opened Thu Jan 30 00:00:41 2020 |
| 00:25 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:10 | <@celticminstrel> | XD |
| 01:10 | <@celticminstrel> | Seven years!? |
| 01:30 | < Pinkhair> | Well, at least that is something you can just look up now. :D |
| 01:32 | <@Alek> | now people are looking for pi to the trillion, ne? |
| 01:34 | < Pinkhair> | They've passed two trillion this year |
| 01:35 | < Pinkhair> | er, last year now I guess |
| 01:35 | <@Alek> | cool. |
| 01:36 | <&McMartin> | "In March 2019 Emma Haruka Iwao, a Google employee from Japan, calculated to a new world record length of 31 trillion digits with the help of the company's cloud computing service" |
| 01:36 | <@Reiv> | good times |
| 01:36 | <&McMartin> | https://www.bbc.co.uk/news/technology-47524760 |
| 01:36 | <@Reiv> | Remind me what degree of pi do we need to keep plank length precision in the entire galaxy? |
| 01:37 | <@Reiv> | ah, no |
| 01:37 | <@Reiv> | "To the width of a hydrogen atom in the entire universe" is 39 digits |
| 01:40 | <@Reiv> | Conviniently, this can be stored in 128 bits, if you're willing to store/calculate the decimal place elsewhere. |
| 01:41 | <@Reiv> | Any value beyond this is purely for the hell of it |
| 01:41 | <@Reiv> | but hey, why not |
| 01:41 | <@celticminstrel> | Isn't that rougly equivalent to saying it fits in a 128-bit mantissa…? |
| 01:41 | <@celticminstrel> | And you still need to store the exponent? |
| 01:42 | <&McMartin> | You know the exponent. |
| 01:42 | <@Reiv> | Like I say, you need to track the exponent elsewhere, but in this case the exponent is always going to be fixed in value, so |
| 01:42 | <&McMartin> | Also the sign, for that matter~ |
| 01:42 | <@Reiv> | Because you are not storing an arbitary value here, but the values of pi you need |
| 01:42 | <@celticminstrel> | True, if it's pi the exponent is known. sure. |
| 01:43 | <@Reiv> | Which is why I'm willing to note it as not strictly part of the number, or at least, not a part of the number that cannot be trivially dealt with while converting the number above to something you can, y'know, work with in actual code |
| 01:43 | <@Reiv> | (Shoving it into a floating point would rather defeat the purpose, after all) |
| 01:44 | <@Reiv> | (Though now I am curious just how many bits of floating point you need to be able to accurately define that) |
| 01:44 | <&McMartin> | Oh hey, there's an implementation of the fastest algo for this in Python (which has unlimited precision rationals) on Wiki |
| 01:44 | <&McMartin> | https://en.wikipedia.org/wiki/Chudnovsky_algorithm |
| 01:44 | <&McMartin> | Unlimited-precision rational is probably the way to go here over any kind of IEEE 754-like floating point system |
| 01:45 | <@celticminstrel> | IEEE doesn't even define anything larger than 128 bits. |
| 01:45 | <&McMartin> | That said, as you'd expect, *in* IEEE 754 128-bit isn't quite enough to hold it right |
| 01:45 | <&McMartin> | I see an Octuple-Precision table here on Wiki. |
| 01:45 | <@celticminstrel> | Oh, really? When did that appear? |
| 01:45 | <&McMartin> | Though it also notes that it is an interchange format, not a "basic" one. |
| 01:46 | <@celticminstrel> | I guess that would be 256 bits. |
| 01:46 | <&McMartin> | Yep |
| 01:46 | <@celticminstrel> | Which is probably a lot more than the minimum needed for this representation of pi. |
| 01:46 | <&McMartin> | 1 sign bit, 19 exponent bits, 236 bits of mantissa |
| 01:46 | <&McMartin> | (237, but you never store the first 1) |
| 01:47 | <&McMartin> | Looks like it was added in 2008. |
| 01:47 | <@celticminstrel> | Is there like, a formule to define an IEEE-style number in any given number of bytes? |
| 01:47 | <@Reiv> | McMartin: Unlimited precision rationals being the ones where you just type out the number then note where the exponent goes right |
| 01:47 | <@Reiv> | ah yes |
| 01:47 | <@Reiv> | You know you just taught me a thing with that line there McM |
| 01:48 | <@Reiv> | I always wondered why so /many/ of our columns in our Oracle codebase at the old job were set to Magnitude 19 |
| 01:48 | <@Reiv> | Like, it was sure as hell /enough/, we were tracking mere dollar values in them like terrible people |
| 01:48 | <@Reiv> | But I did wonder why it was /nineteen/ |
| 01:50 | <&McMartin> | I would hesitate to conclude that based on this :) |
| 01:51 | <@Reiv> | I would not! |
| 01:51 | <@Reiv> | But that is because I knew the devs had a habit of being very precise in their laziness~ |
| 01:52 | <@Reiv> | We'd see a /lot/ of, uh, default data structures |
| 01:53 | <@Reiv> | This wasn't inherently bad |
| 01:53 | <@Reiv> | But it was noticable |
| 01:53 | <@Reiv> | Once you knew what you were looking for, as it were |
| 01:53 | <@Reiv> | Normally you wouldn't think twice about it, but in databases 'data type' is a huge deal so you're always checking, and that's when it starts to stand out |
| 01:58 | <&McMartin> | So, the reason I would be skeptical is because The Laziest is 11-bit. |
| 01:59 | <&McMartin> | That's "double precision", the 64-bit standard that has been hardware-standard since like 1985 and still is today |
| 02:00 | <&McMartin> | (it is also the one that, e.g. Perl, Java, and JavaScript have specified in their specifications, with a handful of caveats) |
| 02:02 | <@Reiv> | yeah no I hn |
| 02:02 | <@Reiv> | They avoided floating points as much as possible in the system |
| 02:02 | <@Reiv> | like the plague |
| 02:02 | <@Reiv> | also note this is a database |
| 02:09 | <@Reiv> | Oracle FLOAT is a subtype of NUMBER, and NUMBER does the 19 decimals of precision thing if you let it by default |
| 02:09 | <@Reiv> | Which is fine but you really only needed, like, three, and would have been more efficient in space if you did, but whatevs, I guess |
| 02:10 | <@Reiv> | (Databases: The only time in modern programming you should probably still be worrying about how many bits your data fits into.) |
| 02:10 | <&McMartin> | OK, that is coincidental. |
| 02:11 | <&McMartin> | Because the 19 up there isn't "19 decimals of precision" |
| 02:11 | <&McMartin> | It's "when writing in scientific notation, you can use exponents up to about half a million" |
| 02:12 | <@Reiv> | oh |
| 02:12 | <@Reiv> | lol, alrighty then! |
| 02:12 | <@Reiv> | oh, yes, it comes back to me now |
| 02:13 | <@Reiv> | You had NUMBER(Precision,scale) |
| 02:13 | <@Reiv> | Scale seemed to eternally be set to 19, which meant a number could be stored up to 19decimal places |
| 02:13 | <@Reiv> | Which was awfully specific to be a default, but hey ho |
| 03:58 | | Degi [Degi@Nightstar-b8c30a.dyn.telefonica.de] has quit [Operation timed out] |
| 04:01 | | Degi [Degi@Nightstar-hcpcro.dyn.telefonica.de] has joined #code |
| 05:00 | | macdjord|dance is now known as macdjord|slep |
| 05:03 | | Vorntastic [uid293981@Nightstar-h2b233.irccloud.com] has joined #code |
| 05:03 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 05:22 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Ping timeout: 121 seconds] |
| 05:43 | | Derakon is now known as Derakon[AFK] |
| 05:59 | | celticminstrel [celticminst@Nightstar-0p84vo.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:05 | | SmithKurosaki [uid215460@Nightstar-ed0oqj.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 08:16 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 08:16 | | mode/#code [+o himi] by ChanServ |
| 10:22 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 10:22 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 10:22 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 10:49 | <@sshine> | I just found an 18 year old logic bug in the code. |
| 10:49 | <@sshine> | at work, that is. |
| 10:49 | <@sshine> | return 2 unless @$l1 > 0 and @$l2 > 0; |
| 10:49 | <@sshine> | "and" is like "&&" except it has a very low precedence. so that reads as: (return 2 unless @$l1 > 0) and @$l2 > 0; |
| 10:52 | <@sshine> | so, just in case the universe wanted to know if @$l2 was empty, it gets computed once in a while. |
| 10:53 | <~Vorntastic> | Neat |
| 10:53 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 10:54 | <@sshine> | this is about the most unserious piece of software I've spent time on. |
| 10:55 | <@sshine> | I should get this done now. |
| 10:58 | <@sshine> | comment above this function is excellent scare tactic: When the metric for a list is calculated we use a technique from vector calculus [...] |
| 11:33 | < Emmy> | eeek |
| 11:36 | <@sshine> | yeah, better not challenge the vector calculus gods |
| 11:39 | <@sshine> | interestingly, the way it calculates this metric distance works even though I change the vector entries from integers to strings. |
| 11:39 | <@sshine> | most often changing the type in Perl programs has scary side-effects. |
| 12:16 | | Kindamoody[zZz] is now known as Kindamoody |
| 12:24 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 12:24 | | mode/#code [+o macdjord] by ChanServ |
| 12:26 | | macdjord|slep [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 12:41 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 12:44 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 12:44 | | mode/#code [+o macdjord] by ChanServ |
| 12:48 | | mac [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 12:48 | | mode/#code [+o mac] by ChanServ |
| 12:51 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 13:13 | | macdjord|slep [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 13:13 | | mode/#code [+o macdjord|slep] by ChanServ |
| 13:15 | | mac [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 13:57 | | celticminstrel [celticminst@Nightstar-0p84vo.dsl.bell.ca] has joined #code |
| 13:57 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:10 | | celticminstrel is now known as celmin|away |
| 15:03 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:03 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 16:43 | | Vorntastic [uid293981@Nightstar-h2b233.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 16:52 | | * McMartin once again encounters the open-source designation "free-as-in-puppy" |
| 16:53 | | McMartin [mcmartin@Nightstar-c25omi.ca.comcast.net] has quit [Connection closed] |
| 16:54 | <@TheWatcher> | ... wat |
| 16:56 | < ErikMesoy> | Costs maintenance and attention? |
| 17:00 | <~Vornicus> | Sounds about right |
| 17:29 | | macdjord|slep is now known as macdjord |
| 17:30 | <@macdjord> | TheWatcher: 'We're actively /trying/ to give this away please, we have way too many'? |
| 17:32 | <~Vornicus> | that's free as in scrap. |
| 17:37 | <@macdjord> | Reiv: Re: Pi: "To the width of a hydrogen atom in the entire universe" is only 39 digits if your equation only uses pi multiplicatively. If you have pi^2 as a factor, or worse pi in an exponent, then you might need more accuracy. |
| 17:38 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 18:02 | | mac [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 18:02 | | mode/#code [+o mac] by ChanServ |
| 18:05 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 18:17 | | McMartin [mcmartin@Nightstar-c25omi.ca.comcast.net] has joined #code |
| 18:17 | | mode/#code [+ao McMartin McMartin] by ChanServ |
| 18:36 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 18:36 | | mode/#code [+o macdjord] by ChanServ |
| 18:38 | | mac [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 19:48 | <@Reiv> | how often does pi come up as an exponent in measurement |
| 20:17 | < ErikMesoy> | Reiv: Variations on e^pi, mostly. https://en.wikipedia.org/wiki/Gelfond%27s_constant |
| 20:29 | <@Reiv> | Okay, that's fair |
| 20:29 | <@Reiv> | I mean, it's not needed for /measurement/ at that point, but still has a use in theoretical math, I suppose |
| 20:29 | <@Reiv> | (Unless it turns out we totally use the stuff for obscure physics or the like, but) |
| 20:52 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 21:22 | | Derakon[AFK] is now known as Derakon |
| 22:54 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 22:54 | | mode/#code [+o himi] by ChanServ |
| 23:14 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| 23:46 | | Kindamoody is now known as Kindamoody[zZz] |
| --- Log closed Fri Jan 31 00:00:42 2020 |