UmbralRaptor changed the topic of #kspacademia to: https://gist.github.com/pdn4kd/164b9b85435d87afbec0c3a7e69d3e6d | Dogs are cats. Spiders are cat interferometers. | Космизм сегодня! | Document well, for tomorrow you may get mauled by a ネコバス. | <UmbralRaptor> … one of the other grad students just compared me to nomal O_o | <ferram4> I shall beat my problems to death with an engineer.
Technicalfool has quit [Ping timeout: 201 seconds]
<xShadowx>
does hypnotism accually exist outside of hollywood? everything i try to google is either stupid rigged stuff, or how to hypnotize a girl to love you o.O .....ya....if that worked we'd have world peace :P
<Fiora>
it does exist, but it's nothing like stage hypnotism
<xShadowx>
no quack like a dog bark like a cat?;p
<Fiora>
you've probably been hypnotized countless times throughout your life, by accident, probably by yourself. it's just a mind-state
<SnoopJeDi>
I'm about 2/3 of the way through Goleman's book on Focus, it's been really neat to learn about the accumulated body of knowledge about that sort of mental machinery
<Fiora>
it's a similar state to the one wherein you space out at your desk but you're not asleep
<xShadowx>
nah only by the tv and the double Ds next door at college
<Fiora>
you're technically "awake", and "conscious"
<Fiora>
being "obsessed" with something is if anything quite divorced from the idea of being hypnotized :P
<SnoopJeDi>
Oh are you talking about "flow" Fiora?
<bofh>
(now delivering: fire)
<Fiora>
i kind of only vaguely remember the concepts honestly, last time i read a hypnotherapy textbook was about 5 years ago
<SnoopJeDi>
yea I'm not super familiar with the technical bits, only ever taken one psych course.
<SnoopJeDi>
Enjoying the book quite a lot though, it fits hand-in-glove with his other work
<SnoopJeDi>
Particularly Emotional Intelligence which was a huuuuge deal for me
<bofh>
Oh?
<Fiora>
to note, the book i read also made a fairly significant point of hypnotherapy being just another method, and not categorically that different from other psychotherapy methods (and requiring similar understanding of the person)
<SnoopJeDi>
yea, mentally I consider that book the point I "turned the corner" on emotion and became actually fluent at all
<SnoopJeDi>
he's like the full-sugar version to Malcom Gladwell's diet-popsci, very good endnotes.
<SnoopJeDi>
Fiora, one of my favorite parts of "Red Mars" was the bit when the psychologist character ponders psychology as a whole as being a really squishy mixed-bag approach to trying to talk about brains
<SnoopJeDi>
he was thinking about the sense in which the four humours are kinda applicable or somesuch and Robinson just kinda sticks in this little cheeky jab at psych
<bofh>
Malcolm Gladwell is kind of a raging tire fire IMO tho, especially since it feels like a study out of THINKING, FAST & STRONG gets disproved each week lately.
<SnoopJeDi>
I mean, I *enjoyed* the books by him that I read, but it does seem to be a really casual/lazy form of engagement in whatever subject
<SnoopJeDi>
And Goleman's not perfect either, I have a few post-its marking remarks that suggest a quick fact-check in the reference they primarily come from. But hey even having those references to do that process is a leg up
<bofh>
True.
<bofh>
Also I just realised I mixed up Gladwell and Daniel Kahneman earlier, whoops. My apologies.
<bofh>
Gladwell is pop-psych but nowhere near Kahneman's level of awful.
<SnoopJeDi>
But yea, I nabbed the book at the library because I'd just happened to lay eyes on it while I waited at the circulation desk. I'd had my interest piqued by the book in the past so I figured that was enough coincidence for my taste
<SnoopJeDi>
It's good enough that I think I'd like to own a copy, so I'll probably be replacing my gifted-away copy of Emotional Intelligence and then picking up a copy of this
<SnoopJeDi>
also read "The Three Body Problem" while I was on a mini-hiatus from here and *oh man* that book was fun.
<kmath>
YouTube - Entry for the Flashing light prize 2017
<bofh>
SnoopJeDi: interesting, might give it a read at some point maybe. what's it about?
<SnoopJeDi>
3BP? It's pretty much literally about the eponymous problem, except that spins the plot off into this craziness with SETI/etc.
<bofh>
ooh, okay definetly giving it a read then
<SnoopJeDi>
and a lot of it is backset during the Cultural Revolution with loooots of historical detail so
<SnoopJeDi>
yea, a fun read
<SnoopJeDi>
(it and the two sequels were written in Chinese and translated)
<bofh>
OHHHHHHHHHHHHHHHHH THIS NOVEL
<bofh>
It's been on my to-read list for months.
<bofh>
I think I'll read it (amusingly) on my flight out from PRC
<SnoopJeDi>
yea it's been on my goodreads pretty much since it won a Nebula in 2014
<SnoopJeDi>
I absolutely intend to follow up with the sequels, it scratched an itch for very novel premise in such an amazing way
<SnoopJeDi>
As in you read it and think "jeez this plot would get laughed out of most pitch meetings" but then it's executed so plausibly (with a few pinches of phlebotinum to make it all go)
<bofh>
which has a similar great plot that sounds at first like it'd get laughed out of a pitch meeting.
<SnoopJeDi>
And the historical bits were just neat. The keyhole satellite program is described in some detail, and I hadn't ever stopped to consider that space reconnaissance had film capsules that re-entered the atmo
<bofh>
oh, yep, it did. same for US & RU
<SnoopJeDi>
But like, that's an *insane* thought in modern context. The past really is a foreign country.
<bofh>
like, never underestimate the bandwidth of a kodachrome cylinder accelerating towards earth at ~9.81m/s^2
<SnoopJeDi>
hmm, was Spectrum ever translated?
<bofh>
especially when your alternative was uh... okay so Voyager 1/2 had literally the best quality imaging capability at the time with its vidicons and even that was still inferior res against film in 1977, just it's kinda hard to recover cylinders from Neptune :p
<SnoopJeDi>
I can't find anything, which is a shame :(
<bofh>
and it's a fucking laugh riot
<SnoopJeDi>
but I would love to read a book of Russian origin in the same spirit, just for wanting to get inside the heads of Russian colleagues a bit more
<bofh>
"Научно-исследовательский институт Чародейства и Волшебства. (НИИЧАВО)"
<bofh>
*dies laughing*
<bofh>
(roughly speaking "Scientific Research Institude of Sorcery & Wizardry", which acronyms out to something that sounds identically to how ничего ("nothing") is pronounced)
<SnoopJeDi>
:D
<SnoopJeDi>
I like the way puns work in not-English languages
icefire has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
<SnoopJeDi>
like learning Japanese sounds worth it for the puns alone
<SnoopJeDi>
some of the only Russian I know I learned from a friend of my advisor at a workshop: "не невозможное но очень трудна"
<SnoopJeDi>
and considering how much talk about muon cooling was going on at the workshop, it's a particularly hilarious punch at the sorts of ideas people come up with
regex has quit [Remote host closed the connection]
tfool has joined #kspacademia
Technicalfool_ has quit [Ping timeout: 204 seconds]
UmbralRaptor has quit [Ping timeout: 204 seconds]
UmbralRaptor has joined #kspacademia
<egg|zzz|egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor a Hitachi protractor with a comma attachment
<egg|zzz|egg>
!wpn Fiora
* Qboid
gives Fiora a double thagomizer
egg|zzz|egg is now known as egg
ferram4 has quit [Read error: Connection reset by peer]
ferram4 has joined #kspacademia
ferram4 has quit [Ping timeout: 204 seconds]
<UmbralRaptor>
!wpn egg|work|egg
* Qboid
gives egg|work|egg an impedance-matched demon
<egg|work|egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor an ellied induction with a platypus attachment
<egg|work|egg>
!tell lamont so we're going to start the release process for Cayley tonight/tomorrow, what with the new moon being on the 23rd. That means your Makefile changes won't make it into Cayley; can we try getting them in for Чебышёв (the solar eclipse release)?
<Qboid>
egg|work|egg: I'll redirect this as soon as they are around.
<BPlayer>
Hi! Sorry I left yesterday without warning. I was hanging out on KSPOfficial and announced I was about to leave there, but I totally forgot I was still in kspacademia.
<BPlayer>
lamont, I saved everything you said here up until 19:25 PM UTC.
<egg|work|egg>
BPlayer: there are logs for the channel, so you can see everything that was said at https://logs.tmsp.io/kspacademia/
<egg|work|egg>
(logs are by Thomas using whitequark's software)
<BPlayer>
Wow, this channel has so much more useful stuff than KSPOfficial! :D Thanks a lot!
<UmbralRaptor>
Assuming Legolas has anime eyes, he can plausibly resolve the heads of the riders. Even without, counting the number of riders is completely plausible.
<bofh>
UmbralRaptor: what, are we trying to determine minimum resolvable phase angle from Legolas' eyes?
<SnoopJeDi>
I'm tempted to just rip this fieldmap nonsense out entirely and hodgepodge better HDF5 support
<SnoopJeDi>
Reading 18 MB of information from a text file should not take O(1 hr)
<bofh>
LOL ARE YOU SHITTING ME
<bofh>
HOW?!?
<Iskierka>
is KSP loading it?
<UmbralRaptor>
bofh: sorta. Blame Iskierka
<SnoopJeDi>
2.1 M calls to fscanf() (and corresponding calls to fprintf())
<UmbralRaptor>
SnoopJeDi: o_O
<bofh>
Ahh, I just saw the parenthetical and what the loops iterate over.
regex has joined #kspacademia
<bofh>
fscanf() is dogshite.
<SnoopJeDi>
at least, I think. I'm not really familiar with how to use valgrind very well
<bofh>
Like it's slow as fuck and there
<SnoopJeDi>
But like even using fscanf wouldn't be so bad if you read the entire file in and then parsed it or whatever?
<bofh>
Oh, that's a good point.
<SnoopJeDi>
my C++ fu isn't great but this is like remarkably poorly designed
<bofh>
Using sscanf in sequence (tho I'd use strtod() in that case)
<bofh>
(OR JUST FUCKING HDF5 LIKE A SANE PERSON)
<SnoopJeDi>
I've been kinda edging towards a fully-HDF5 solution because it has SOME support for its own subsidiary HDF5 format
<SnoopJeDi>
I just want adequately-resolved focusing quadrupoles ;_;
<bofh>
welcome to scientific computing
<SnoopJeDi>
yup
<SnoopJeDi>
I can frame it as a positive, I've effectively added valgrind to my I-dont-really-know-this-tool-but-I-know-vaguely-of-it toolbelt
<Fiora>
HDF5......
<Fiora>
....... sane..........
<SnoopJeDi>
shades of relativity
<bofh>
^
<SnoopJeDi>
HDF5 is downright rational compared to "hey don't put too many (or too few) decimal places or this will fail but tell you it succeeded"
<bofh>
compared to $ad_hoc_ascii_floating_point_table format, I'd say it's sane.
<SnoopJeDi>
at least I'm not an astrophysicist I guess
<bofh>
OH MAN that reminds me of the input format to a compchem app before I got through with it a few years ago
<bofh>
That one was fun, it was "don't put too many/too few LEADING SPACES or this will fail or read utter gibberish but tell you it succeeded"
<SnoopJeDi>
bofh, so I'm running a simulation I know will fail just so that I can *hope* valgrind captures something useful when it decides to try and read the fieldmap *again*
<bofh>
I too love fixed-width formatted Fortran input strings.
<SnoopJeDi>
because that's what it does: it says "fieldmap read successfully!" and then immediately re-initializes my cyclotron element
<SnoopJeDi>
causing it to read the fieldmap again -_-
<bofh>
what the fuck
<SnoopJeDi>
so it's not even as good as "wait an hour or two and then the tracker will run" it's "you're fucked if you want a large, hi-res map" as the tool is written
<SnoopJeDi>
or at least if you want that from this particular file format. It supports like 7 and the user manual explicitly tells you to just cram your own shit in if you feel like it
<bofh>
now I know why it's such a disasterfuck.
<bofh>
"cram your own fileformat in" is never a sane approach to things.
<SnoopJeDi>
Yea the word "horizon" is apparently alien to this group
<SnoopJeDi>
I would hazard a guess 20-30% of the codebase is unnecessary duplication, at least
<SnoopJeDi>
hard to tell when methods are regularly ballooned over 1000 lines
<SnoopJeDi>
you wrote all that shit, and I ain't reading it
<bofh>
ahh, sounds like scientific computing
<SnoopJeDi>
sure 'nuff
<bofh>
I love how this shit is the same tire fire everywhere
<SnoopJeDi>
I wonder if it'll ever matter, sometimes
<SnoopJeDi>
like, we've lost so much work from the 90s and nobody really seems to care, so maybe this is gonna be the same thing
<bofh>
¯\_(ツ)_/¯
<SnoopJeDi>
data shepherding, reproducibility, blah blah, doesn't seem like any of it matters
<bofh>
a lot depends on the dataset IMO
<SnoopJeDi>
I lost hope pretty much entirely when I read about Lorena Barba's group trying to replicate their own work
<bofh>
Oh?
<SnoopJeDi>
yea, it's a very depressing read: they followed every best practice, e.g. VCS, good documentation
<SnoopJeDi>
Couldn't do it
<SnoopJeDi>
In fairness, it was CFD so that's not a hard task to fail at
<SnoopJeDi>
but still you'd think in a just world, running *exactly* the same input would give you at least comparable output and it totally doesn't
<bofh>
what the fuck
<SnoopJeDi>
At one point I felt pretty good about containerization as a way to address the problems that paper raises but I've mostly lost that zeal. At this point I'm just doing it for *my* work as a point of personal organization
<SnoopJeDi>
Oh huh, interesting. I let the damned thing run for an hour and a half (!) until it failed-over, and on that timescale the fprintf() is way more dominant
<SnoopJeDi>
45% in fprintf(), only around 7% in fscanf()
<bofh>
what the everliving fuck is it even printing?
<SnoopJeDi>
bofh, so it turns out that write-out behavior is an "info" feature that just comes enabled by default, at least there's a flag to disable it (and like 10,000 other potentially useful things)
<SnoopJeDi>
I feel slightly less petty about that particular garbage fire because it doesn't require me to lobotomize the loading routine and run my own build of the program.
<SnoopJeDi>
(it's actually pretty zippy without it, fscanf() notwithstanding)
<bofh>
SnoopJeDi: ...I'm actually surprised *turning off the fucking debug prints, which were inexplicably on by default* sped it up that dramatically.
<bofh>
!wpn egg
* Qboid
gives egg an oncological furious-sounding thermometer
<SnoopJeDi>
Well it wasn't printing to the screen at least but yea it's kinda surprising to me how big the difference between the read and write is
<bofh>
SnoopJeDi: but hey, not needing to maintain your own fork just yet is nice.
<SnoopJeDi>
heh, well-placed "yet"
<SnoopJeDi>
having faster runs is nice, I'm not terribly good at juggling an on-going simulation with other work
<bofh>
Heh.
<bofh>
Hey, I'm just calling 'em as I've seen 'em. Distressingly many times before. :P
<SnoopJeDi>
Yea I accepted it as a foregone conclusion that my thesis will probably run on some Frankenstein's-monster version of this codebase :)
<bofh>
my condolences.
regex has quit [Remote host closed the connection]
ferram4 has quit [Read error: Connection reset by peer]
ferram4 has joined #kspacademia
* egg
had fallen asleep for some reason
egg is now known as egg|zzz|egg
<egg|zzz|egg>
!wpn bofh
* Qboid
gives bofh a honed graviton rotor
<UmbralRaptor>
"In some situations, the optimizer replaces certain numbers in the code with routines that compute different numbers"
<UmbralRaptor>
AAAAAAAAAAAAAAAAA
<egg|zzz|egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor an iodine cryocooler
<UmbralRaptor>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a ferromagnetic icosahedron which strongly resembles a commutator