Post AT2Qfgj2PAdw27upaC by niconiconi@mk.absturztau.be
(DIR) More posts by niconiconi@mk.absturztau.be
(DIR) Post #ASzuozNepTWreFO1dQ by niconiconi@mk.absturztau.be
2023-02-24T07:35:39.378Z
10 likes, 13 repeats
FORTRAN to WebAssembly cross-compile is a success! I can run FORTRAN code in Firefox now!!!
(DIR) Post #ASzuw0DQ7ioGGSwUZU by puniko@mk.absturztau.be
2023-02-24T07:36:56.587Z
1 likes, 0 repeats
@niconiconi :blobcataww: you are heccing AWESOME
(DIR) Post #ASzv7lhltnFMjhkWae by fink@chaos.social
2023-02-24T07:36:59Z
0 likes, 1 repeats
@niconiconi fortfox? Firetran?
(DIR) Post #ASzvnkWmfM7BFPd0Lo by w@arachnid.town
2023-02-24T07:46:03.237245Z
1 likes, 0 repeats
@niconiconi
(DIR) Post #ASzvrFJ8gR8bZvufi4 by maddiefuzz@hackers.town
2023-02-24T07:46:31Z
1 likes, 0 repeats
@puniko @niconiconi I’ve been excited about this happening, it’s so cool
(DIR) Post #AT00cQd6o0FI7X0X9U by roytam1@miniwa.moe
2023-02-24T08:40:38.184926Z
0 likes, 0 repeats
@niconiconi next: use w2c2 to convert it to C89.https://github.com/turbolent/w2c2
(DIR) Post #AT282zPtBCCrXBM9ui by arclight@oldbytes.space
2023-02-25T08:53:28Z
1 likes, 0 repeats
@niconiconi Curious what it would take to port Castlequest to WebAssembly https://gitlab.com/apthorpe/CastlequestI modernized the F66 original to F2018 and in the process removed the ASA carriage control codes so it plays well on modern terminals. The gameplay should be identical though it recognizes lower case now and I fixed a few crashes and ambiguous behaviour.Anyway, you deserve some fun after all that work :)
(DIR) Post #AT2BbKen0yFj7OOSGG by koen_hufkens@mastodon.social
2023-02-25T09:52:47Z
0 likes, 1 repeats
@niconiconi Time to run some climate models in a browser tab!
(DIR) Post #AT2BknjiGdyZueDAvo by niconiconi@mk.absturztau.be
2023-02-25T09:54:45.965Z
0 likes, 0 repeats
@koen_hufkens@mastodon.social A real possibility. I haven't tried it yet, but according to even earlier attempts from previous experimenters, LAPACK and BLAS worked just fine.
(DIR) Post #AT2Ntxs79UdtP5HF3Y by loke@functional.cafe
2023-02-25T12:07:58Z
0 likes, 0 repeats
@niconiconi OK, computing is done. No need to to do anything else. It's all downhill from here I guess.
(DIR) Post #AT2NtyWWjEy7QQBW2y by niconiconi@mk.absturztau.be
2023-02-25T12:10:52.555Z
0 likes, 0 repeats
@loke@functional.cafe If I tell you that I've just used this technique to port an entire boundary-element method electromagnetic field solver to a browser tab... The solver was from the 90s and some FORTRAN code was from the 70s...
(DIR) Post #AT2Qfg7odYrwAgV6Z6 by jamiemccarthy@mastodon.social
2023-02-25T12:37:29Z
0 likes, 0 repeats
@niconiconi That’s amazing and wrong. The 8th decimal digit is off by one. No idea how that algorithm got back on track for the following digits.
(DIR) Post #AT2Qfgj2PAdw27upaC by niconiconi@mk.absturztau.be
2023-02-25T12:41:56.946Z
0 likes, 0 repeats
@jamiemccarthy@mastodon.social It's NOT wrong within the limitation of the algorithm, my understanding is that the simple Leibniz formula is just incapable of achieving a higher precision, even using a huge number of terms. Speaking of why the 8th digit is wrong but the 9th digit is right, I think it's just a coincidence. Leibniz formula doesn't calculate π but π/4, so perhaps it's related.
(DIR) Post #AT2QfhmyS1NzKclUMC by jamiemccarthy@mastodon.social
2023-02-25T12:40:02Z
0 likes, 0 repeats
@niconiconi Unless that’s just where the precision ends and it’s total coincidence that the following 4 digits are correct, but wouldn’t that be a 1 in 10,000 shot? Hm
(DIR) Post #AT2QlAfaV3uCLsusPQ by niconiconi@mk.absturztau.be
2023-02-25T12:42:57.668Z
0 likes, 0 repeats
@jamiemccarthy@mastodon.social It's NOT wrong within the limitation of the algorithm, it's just a simple series expansion. My understanding is that the Leibniz formula is just incapable of achieving a higher precision even with a huge number of terms. Speaking of why the 8th digit is wrong but the 9th digit is right, I think it's just a coincidence. Leibniz formula doesn't calculate π but π/4, so perhaps it's related.
(DIR) Post #AT2QsF72Ne7A9r7z0K by niconiconi@mk.absturztau.be
2023-02-25T12:44:13.694Z
1 likes, 0 repeats
@jamiemccarthy@mastodon.social For your reference:Leibniz's formula converges extremely slowly: it exhibits sublinear convergence. Calculating π to 10 correct decimal places using direct summation of the series requires precisely five billion terms
(DIR) Post #AT2RMaeOrnSVe2Cgee by niconiconi@mk.absturztau.be
2023-02-25T12:49:43.014Z
0 likes, 0 repeats
@jamiemccarthy@mastodon.social Wikipedia also explains the strange phenomenon of how a truncated series summation can produce somewhat correct digits. Amazingly it's a special mathematical property of this series, and one can even exploit this property to improve the algorithm with accelerated convergence."If the series is truncated at the right time, the decimal expansion of the approximation will agree with that of π for many more digits, except for isolated digits or digit groups. [...] The errors can in fact be predicted; they are generated by the Euler numbers En"
(DIR) Post #AT2RXBY4XHm6yfIPVQ by niconiconi@mk.absturztau.be
2023-02-25T12:51:17.161Z
1 likes, 0 repeats
@jamiemccarthy@mastodon.social It turned out not to be a coincidence but an inherent property of the series. One can even exploit this property to improve the algorithm with accelerated convergence."If the series is truncated at the right time, the decimal expansion of the approximation will agree with that of π for many more digits, except for isolated digits or digit groups. [...] The errors can in fact be predicted; they are generated by the Euler numbers En" https://en.wikipedia.org/wiki/Leibniz_formula_for_%CF%80#Unusual_behaviour
(DIR) Post #AT2SvtamXAGeTruU6a by mikemol@pony.social
2023-02-25T13:05:46Z
0 likes, 0 repeats
@niconiconi Does this mean we'll have more authentic bank simulator games?
(DIR) Post #AT2SvuCMHSKEMPUUfw by niconiconi@mk.absturztau.be
2023-02-25T13:07:15.200Z
0 likes, 0 repeats
@mikemol@pony.social Aren't those powered by COBOL?
(DIR) Post #AT2TIY6YaduuUzZgYa by mikemol@pony.social
2023-02-25T13:07:59Z
0 likes, 0 repeats
@niconiconi Hm. You're right. Too early in the morning here.
(DIR) Post #AT2TIYZypFS7yFLBQ0 by niconiconi@mk.absturztau.be
2023-02-25T13:11:21.710Z
0 likes, 0 repeats
@mikemol@pony.social "Fortran is for wimpy engineers who wear white socks, pipe stress freaks, and crystallography weenies. They get excited over finite state analysis and nuclear reactor simulation."
(DIR) Post #AT3vhKxRbCGQq2sdWq by Humpleupagus@eveningzoo.club
2023-02-26T06:04:19.967331Z
2 likes, 0 repeats
I thought Firefox was already for trans. 🤔
(DIR) Post #AT3viqhP4THnZW69Zo by theorytoe@ak.kyaruc.moe
2023-02-26T06:04:37.535431Z
2 likes, 1 repeats
@niconiconi >FORTRAN>runs on the webtruly FORTRAN has literally outlived any other language LMAO
(DIR) Post #AT3vwkNhMDi221VaXQ by niconiconi@mk.absturztau.be
2023-02-26T06:07:05.375Z
1 likes, 0 repeats
@theorytoe@ak.kyaruc.moe True. At this point I've already executed a (tiny) portion of the LINPACK code in FORTRAN 77 via Node.js, it worked just fine and solved my matrix. This is the original LINPACK from 1978!
(DIR) Post #AT3w12bNYUhlhTxGnA by theorytoe@ak.kyaruc.moe
2023-02-26T06:07:54.854655Z
0 likes, 0 repeats
@niconiconi DAMN :gyate_hijiri:
(DIR) Post #AT3w6423fHpvaHpT3Q by ceo_of_monoeye_dating@bae.st
2023-02-26T06:08:49.089042Z
4 likes, 3 repeats
@niconiconi >Making anything resembling an incentive for anyone to make new FORTRAN code everdue_to_personal_reasons_you_will_be_passing_away.jpg
(DIR) Post #AT3wAFWqrXMUOnEDVg by niconiconi@mk.absturztau.be
2023-02-26T06:09:32.259Z
0 likes, 0 repeats
@ceo_of_monoeye_dating@bae.st no, not new fortran code... Just to keep the old one alive. :blobcatevil:
(DIR) Post #AT3wCN3JrQAeJQQUnw by ceo_of_monoeye_dating@bae.st
2023-02-26T06:09:57.818983Z
1 likes, 0 repeats
@niconiconi You know that's not what's going to happen, just like I do.
(DIR) Post #AT3wmKA85F3Dm1kuNk by niconiconi@mk.absturztau.be
2023-02-26T06:16:25.577Z
1 likes, 0 repeats
@Humpleupagus@eveningzoo.club Q: What is the best CPU for viewing Federated Timeline on the Fediverse?A: Transmeta.
(DIR) Post #AT7sHnCgR8SrgKyZCi by buercher@tooting.ch
2023-02-27T21:49:40Z
0 likes, 0 repeats
@niconiconi @hrbrmstr Shouldn’t pi be 3.1415926535…
(DIR) Post #AT7sHniwVCGjIO4KUC by hrbrmstr@mastodon.social
2023-02-27T21:52:24Z
0 likes, 0 repeats
@buercher @niconiconi that's (a) a good catch and (b) kinda scary
(DIR) Post #AT7sHopMOozqia4y80 by niconiconi@mk.absturztau.be
2023-02-28T03:44:53.906Z
0 likes, 0 repeats
@hrbrmstr@mastodon.social @buercher@tooting.ch Leibniz's formula is only a simple approximation based on an incredibly inefficient series expansion, so of course not all digits is going to be correct.
(DIR) Post #AT7sa0ZlSQl8MrSRQO by niconiconi@mk.absturztau.be
2023-02-28T03:48:13.920Z
0 likes, 0 repeats
@hrbrmstr@mastodon.social @buercher@tooting.ch It's not a bug, it's a feature. Leibniz's formula is only a simple approximation based on an incredibly inefficient series expansion, so of course not all digits are going to be correct. Leibniz's formula also has a mathematical quirk for generating the wrong digits between correct digits, so it proves the code is correct.
(DIR) Post #ATAOTyG2kUcCI3JmzY by moffintosh@berserker.town
2023-03-01T08:55:08Z
0 likes, 0 repeats
@niconiconi whyyyyyyyyy