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.
Qboid was kicked from #kspacademia by *status [You have been disconnected from the IRC server]
Qboid has joined #kspacademia
Orkeren has quit [Quit: No Ping reply in 180 seconds.]
Orkeren has joined #kspacademia
e_14159 has quit [Ping timeout: 186 seconds]
e_14159 has joined #kspacademia
<Ellied>
jesus. The spectrometer itself has a nice verbose serial language that makes it pretty easy to figure out what you're doing, but the box that's supposed to control it and the PMT that came with it is *incredibly* terse
<egg|zzz|egg>
UmbralRaptor: funny, but I'm not sure this has any practical use; this is probably rather slower, and good luck doing something higher order than leapfrog
<UmbralRaptor>
SnoopJeDi: so, Jupyter for js visualizations?
<UmbralRaptor>
Iskierka: silly dirb…
<SnoopJeDi>
UmbralRaptor, Kinda™, the ideas he's espousing there go a little deeper. In particular, the reactivity is *very* different from Jupyter, where execution order matters
<SnoopJeDi>
what he's describing slots somewhere between traditional interactive programming (Jupyter/IPython et al.) and declarative visualization frameworks where you're locked into this-or-that pattern
<SnoopJeDi>
I get the impression from the reaction of soundnfury/Sarbianthat reactive programming is a much older idea that's "been done" (e.g. Excel is reactive)
<soundnfury>
SnoopJeDi: part of my reaction was certainly due to "that sounds too much like excel ugh"
<SnoopJeDi>
soundnfury, when I realized exactly how much science is powered almost exclusively by Excel, I caught a sad
<egg|zzz|egg>
and if you do fuzzy equality you will be stabbed by a cat
<SnoopJeDi>
Particularly since it's usually *really* gross Excel that doesn't even use named references etc.
<SnoopJeDi>
I think there was actually a study referenced by Lorena Barba in her reproducibility paper that called out a genetics spreadsheet error as a massive flaw in methodology
<soundnfury>
SnoopJeDi: s/science/business
<Qboid>
soundnfury thinks SnoopJeDi meant to say: soundnfury, when I realized exactly how much business is powered almost exclusively by Excel, I caught a sad
<soundnfury>
and you have why I'm sad too
<SnoopJeDi>
Well it's a lot more intuitive for business
<SnoopJeDi>
And I'd already been exposed to *that* at my job
<SnoopJeDi>
One of the masters students my ops manager hired to do analytics just plopped an Excel sheet with raw data in front of him one day, when asked to look at warehouse shipping patterns or somesuch
<kmath>
YouTube - You Suck at Excel with Joel Spolsky
<egg|zzz|egg>
!wpn whitequark
* Qboid
gives whitequark a proactinium cow
<SnoopJeDi>
but named references are one that I *really* wish more people knew about because it makes formulae much more readable
<bofh_>
you mean readable at all really
<SnoopJeDi>
true enough
<SnoopJeDi>
And things like using $ to 'lock' row/column references independently, blah blah
<SnoopJeDi>
there's a lot under the hood but very few people are curious enough to look, so there's a lot of ugly kludgey VLOOKUP-type stuff in the wild
<whitequark>
I know about $, naturally
<SnoopJeDi>
I think most people know it has something to do with holding, but not that it can be applied separately to rows/cols
<UmbralRaptor>
!wpn egg|notzzz|egg
* Qboid
gives egg|notzzz|egg a boiled submersion
* UmbralRaptor
has no idea how people would do stuff in Excel without $
<SnoopJeDi>
A lot of the first couple of weeks of physics lab teaching ends up being "here's how you do [basic task] with Excel guys"
<UmbralRaptor>
Have a bit of that in my astro lab.
<SnoopJeDi>
My boss hauls out this absolutely atrocious non-linear magnetostatics solver implemented in Excel from time to time
<SnoopJeDi>
where you define the iron boundary and other BCs in terms of marked cells and then let it iterate...
<soundnfury>
I was just about to say that my favourite bit is where I declare the nonexistence of parabolæ
<egg>
it's going to be terribly-conditioned *around e = 1*
<egg>
that's the realy concern
<egg>
nobody gives a damn about parabolae
<SnoopJeDi>
punt early, punt often?
<egg>
but if it doesn't work for them, it ain't gonna work around it
<soundnfury>
well sure, but I had trouble enough finding _any_ equations I could get to work, let alone well-conditioned ones
<egg>
also I went from Conan Doyle citations to deliberate americanisms in 5 lines >_>
<egg>
soundnfury: but that's where all the fun lieth!
<egg>
ok, why am I NaNing my elements
<soundnfury>
basically, after two days of orbital mechanics, I'd had as much as I could take, I wasn't going to turn my nose up at code that, in the majority of cases, works well enough for my porpoises
<soundnfury>
(I realise that in your case, the mathematics and numerical stabbity and so on _is_ the porpoise... but that didn't apply to me)
<egg>
but e.g. if you're dealing with long-period comets this will all break down with great hilarity
<egg>
mind you the code I'm writing right now leaves the conditioning as a TODO
<egg>
partly because it has 15 branches at the widest, each with 5 different tricky equations
<egg>
and for all of those you have to think around the following singularities: parabola, circle, straight ellipse, straight hyperbola
<egg>
AAAAAAA
<soundnfury>
straight ellipse?
<egg>
(and there's also the equatorial case where you need to get fancy with the longitude of periapsis
<egg>
a constant, e -> 1
<egg>
near there the true anomaly is a terrible thing of course