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.
<Ellied>
yours could easily be different.
<APlayer>
I always assumed those had some sort of button or something that makes them wake up when connected?
<Ellied>
I've seen both kinds. Actually, I had one (still do, somewhere, in several pieces) that has an actual two-position on/off switch.
<APlayer>
Heh, sounds fun
<Ellied>
really tiny thing. Barely runs an RPi for an hour, without a screen.
<APlayer>
Okay, I think I understand that resistor too. Another difference between transistors and MOSFETs, it seems
<APlayer>
I don't quite understand the exact numbers, but that's a task for a later time
<APlayer>
For right now it is 2 AM and way too early for such things, actually :P
<Ellied>
yeah. Basically what that resistor does (in some ways) is make the MOSFET act more like a bipolar transistor. The major difference is that you do actually have to supply enough voltage to turn it on, rather than just supplying current through a resistor at any voltage you please.
<Ellied>
if you are driving the MOSFET gate with something that can both source and sink current, you don't need that resistor. The other MOSFET only sinks it, so it's needed in this case.
<APlayer>
Okay then, thanks another pair of huge lots!
<APlayer>
I will run some experiments with materials I have on hand, although I doubt I do have enough. Perhaps I'll take that to school and ask around if any teachers have the parts.
<APlayer>
If it works as expected, I'll buy the parts and solder that circuit, and that will totally be awesome!
<Ellied>
:D
<Ellied>
Tell you what - If you have trouble getting ahold of the parts, I will look into mailing you some. I have *lots* of these things lying around.
<APlayer>
And I am off to get some sleep today at last. Thanks again, and have a nice evening (I believe)!
<Ellied>
sounds good, sleep well!
<APlayer>
That offer is very generous, but I am afraid that shipping from the US to EU will cost more than buying it with free shipping from China. :P Thanks a lot anyway, though!
<APlayer>
And now I am off for real. ;-) Bye!
awang has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
awang has joined #kspacademia
* UmbralRaptor
feels increasingly underprepared mathematically for GMU
<UmbralRaptor>
Of course, 3 weeks in a row of colloquia featuring mathematicians doesn't help. <_<
<UmbralRaptor>
Or "oh, hey. We're just going to introduce the Calculus of variations" I'm classical.
APlayer has quit [Ping timeout: 383 seconds]
e_14159 has quit [Ping timeout: 183 seconds]
e_14159 has joined #kspacademia
icefire has quit [Remote host closed the connection]
<SnoopJeDi>
IME if you're not a heavy-theorist UmbralRaptor it pays massive dividends to admit and lean into the theme of "Boy, I don't understand the mathematics as well as I could"
<SnoopJeDi>
because once you get *past* that, you can start learning it, rather than it being a Big Scary Presence™ in your thoughts (these can be a killer in grad)
<Ellied>
I'm already finding that one
<Ellied>
other student: so then this over this and this times this equals the derivative of all of this, which means this. Me: okay I can see that you understand it quite well but I still don't
<UmbralRaptor>
It's a combination of breadth (3/3 colloquia have been math) and depth (see: relearning partial derivatives. Also that the inner/outer products tutorial egg|zzz|egg gave didn't really get into calculating them, just defining. So for the homework, I went by components)
<SnoopJeDi>
Ellied, it's the frustrating nature of epistemology: it seems so obvious/clear in hindsight.
<Ellied>
yuup
* UmbralRaptor
stares at an attempt to write up a talk about an Apollonian Window.
<Ellied>
I also get trouble a lot with "You were just teaching me electronics that I know nothing about and clearly understand it very well, so how is this math that makes perfect sense to me hard for you?"
<SnoopJeDi>
"Ah yes because indeed I am *learning* it"
<egg|zzz|egg>
bofh: should I tweet it :-p
<egg|zzz|egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor an underclocked octapod McShipFace
<soundnfury>
Ellied: everything is either trivial or impossible
<soundnfury>
!wpn Ellied
* Qboid
gives Ellied a sarcastic instability
<UmbralRaptor>
Major frustration with math/physics: Spend huge amounts of effort building a nifty framework, calculate 0-2 things before moving on to another subject. >_>
<egg|zzz|egg>
every group is virtually trivial or infinite
<SnoopJeDi>
egg|zzz|egg, s/group/thing/
<UmbralRaptor>
soundnfury: It's endlessly frustrating when that happens.
<Qboid>
SnoopJeDi thinks egg|zzz|egg meant to say: every thing is virtually trivial or infinite
<UmbralRaptor>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a ham caffeinated collapsar
<SnoopJeDi>
TIL, of course that's a thing
<UmbralRaptor>
(wait, did I wake egg?)
<egg|zzz|egg>
no but you just made my phone fall off the table
<soundnfury>
UmbralRaptor: I take it you aren't able enjoy the nifty framework for its own niftitude _sans_ applications?
<SnoopJeDi>
elegance: all that is necessary for science?"
<SnoopJeDi>
an ongoing investigative series by apparently some amount of people
* UmbralRaptor
pokes the ratios of the forces with elegance.
<egg|zzz|egg>
UmbralRaptor: tbh a lot of that is that there are many nifty *and also necessary* frameworks to teach, and they take a while to teach, so a bunch of classes are just that; doing things with them happens either in advanced classes or in papers
<bofh>
egg|zzz|egg: yes, yes you should
<SnoopJeDi>
I'm pretty firmly in the "if you can't readily reduce it to vernacular, you're missing something in your understanding" camp
<soundnfury>
SnoopJeDi: all that is necessary to solve for the forces of evil is the minimum of the action of good men
<SnoopJeDi>
("vernacular" is of course dependence on audience)
<SnoopJeDi>
soundnfury, yeesh.
<egg|zzz|egg>
bofh: should I say ⊗ ℂ or ⊗_ℝ ℂ though
<SnoopJeDi>
that's an admiring "yeesh," to clarify
<soundnfury>
phew
<soundnfury>
!wpn SnoopJeDi
* Qboid
gives SnoopJeDi a signaling goto
* egg|zzz|egg
glares at the concept of a signaling goto
<SnoopJeDi>
!wpn soundnfury
* Qboid
gives soundnfury a descriptive Glamdring
<egg|zzz|egg>
!wpn bofh
* Qboid
gives bofh a vegetable automorphism
<soundnfury>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a summer wing
<SnoopJeDi>
egg|zzz|egg, vegetable automorphism...is that an allium thing?
<SnoopJeDi>
or perhaps a brassicaceae thing
<bofh>
egg|zzz|egg: the former seems better, like the latter is more correct mathematically but also extraneous for the joke IMO.
<egg|zzz|egg>
bofh: yeah, also if you have real vector spaces the subscript field is eggstraneous
* egg|zzz|egg
ponders spacing around the equality and tensor product...
<Ellied>
I've often done my best to do explanations xkcd-style with the ten hundred most often used words and such when I explain what I've been up to to my family (building "power things") which they consistently love
<Ellied>
I've tried to do it while TAing too but it's hard to generate it fast enough for a typical lab class
<SnoopJeDi>
I'm a big fan of the "elevator talk" philosophy, that you should have a telescoping series of descriptions so that you can adjust your descriptions (of ideally anything but at least your own work) to various degrees of duration/depth.
<SnoopJeDi>
in no small part due to the success of programs like the 3 Minute Thesis competition and its ilk
* Qboid
gives UmbralRaptor a conjugate diglossic neutronium reactor
<UmbralRaptor>
o_O
<soundnfury>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a solenoidal quill with an elephant attachment
<egg|phone|egg>
zzZ
<egg|phone|egg>
猫⊗猫*
<egg|phone|egg>
Real cat endomorphisms
<UmbralRaptor>
!wa endomorphism
<Qboid>
UmbralRaptor: endomorphism:
<egg|phone|egg>
The Chinese input method is good for mathematics too
<UmbralRaptor>
… thank you, Wolfram. Very helpful.
<UmbralRaptor>
|貓| = 猫?
<egg|phone|egg>
UmbralRaptor: 古在！
* UmbralRaptor
collides with the sun.
APlayer has joined #kspacademia
* UmbralRaptor
? things
* UmbralRaptor
sends a school related email at 2345 on a friday >_<
<bofh>
not sure what's weird about that at all
<UmbralRaptor>
>_>;;
<FluffyFoxeh>
yeah that sounds reasonable to me
<FluffyFoxeh>
:p
<SnoopJeDi>
bit early to be emailing, innit?
<Ellied>
my friends and I are putting together a conspiracy theory that getting a Ph.D. in physics absolves you of the need to eat and/or sleep.
<Ellied>
namely: professors in our department very frequently answer emails in the wee hours of the morning, and I'd worked with my current research prof for over a year before seeing him eat for the first time (and it's only ever been light snacks since)
<Ellied>
also, both times I've been to Argonne (where everyone and their cat has a physics Ph.D.) no one has remembered to feed the undergrads until we badgered them.
<UmbralRaptor>
<_<;;
<Ellied>
also, I don't get the pun.
<FluffyFoxeh>
me neither because I can't read that language
awang has quit [Killed (NickServ (GHOST command used by awang_))]
awang has joined #kspacademia
<egg|phone|egg>
Ellied: cats with doctorates! ?
<egg|phone|egg>
Greening: look up "complexification"
* egg|zzz|egg
starts writing a design doc "10-point plan for the unification of the egg and ferram4 sectors"
<UmbralRaptor>
egg|zzz|egg: Tests will include katabatic winds?
<egg|zzz|egg>
aaaaaaaaaaaaaaaaa
<BPlayer>
Hi!
<soundnfury>
!wpn BPlayer
* Qboid
gives BPlayer a cutting gegenschein-like timer
<BPlayer>
To anyone wondering: I just finished that simulated annealing thing I was working on, and was pleasantly surprised about two things: 1) It works well - much better than expected. And 2) it runs 10 million (not 10 thousand, as initially planned) iterations in the very reasonable time of up to a few minutes, interestingly dependant on the size of the input
<BPlayer>
(Interestingly dependant because I only loop through the input in my fitness function)
<egg|zzz|egg>
!wpn e_
* Qboid
gives e_ an arsenic polynomial
<egg|zzz|egg>
uh
<egg|zzz|egg>
!wpn e_14159
* Qboid
gives e_14159 a Geiger compactification
icefire has quit [Quit: Leaving]
<UmbralRaptor>
!wpn BPlayer && egg|zzz|egg
* Qboid
gives BPlayer && egg|zzz|egg an isenthalpic spectrophotometer
<bofh>
BPlayer: simulated annealing is: A) incredibly fast, B) very useful in practise, often *because* of A)
<bofh>
"This paper gives a straightforward implementation of simulated annealing for solving maximum cut problems and compares its performance to that of some existing heuristic solvers. The formulation used is classical, dating to a 1989 paper of Johnson, Aragon, McGeoch, and Schevon. This implementation uses no structure peculiar to the maximum cut problem, but its low per-iteration cost allows it to find
<bofh>
better solutions than were previously known for ...
<bofh>
... 40 of the 89 standard maximum cut instances tested within a few minutes of computation."
<kmath>
<whitequark> that it's cold and her fur is short, and she wants to continue the walk sitting on my shoulder like a pirate parrot
Technicalfool_ has quit [Ping timeout: 198 seconds]
Technicalfool has joined #kspacademia
<UmbralRaptor>
Shoulder cat!
BPlayer has joined #kspacademia
<egg|zzz|egg>
bofh: well, just make that chemistry game a part of RealFuels :D
<soundnfury>
egg|zzz|egg: not ferram4's realProcEngines?
<soundnfury>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a good balance/Norgg hybrid
<soundnfury>
!wpn -add:adj frumious
<Qboid>
soundnfury: Adjective added!
<ferram4>
I'll implement a combustion solver in ProcEngines at some point and remove the table lookup method.
<ferram4>
But For now, I'm lazy and I really need to refactor everything to separate the combustion chamber, nozzle, turbine, pump, etc. kinda stuff.
<ferram4>
Also calculate masses and some half-ass version of models.
<soundnfury>
!wpn ferram4
* Qboid
gives ferram4 a solvable pharmacy
<BPlayer>
!wpn egg|zzz|egg, soundnfury and ferram4
* Qboid
gives egg|zzz|egg, soundnfury and ferram4 an octanitrocubane ellipsoid
<ferram4>
...that sounds explosive.
<egg|zzz|egg>
ferram4: starting to work on the reunification atm
<UmbralRaptor>
ferram4: shame it lacks the energy to generate a hyperbola.
<UmbralRaptor>
egg|zzz|egg: will the release be named "Charlie?"
<ferram4>
egg|zzz|egg, Aeroschluss? :P
awang has joined #kspacademia
<egg|zzz|egg>
UmbralRaptor: nah, Chasles
<egg|zzz|egg>
ferram4: the forces that you apply are added in the same way as the engine forces, using the relevant method of Part (rather than directly on the rigid body), right?
<egg|zzz|egg>
hm, principia doesn't seem to be picking up your forces
<egg|zzz|egg>
aaaaargh they're add directly to the rbs? D:
<egg|zzz|egg>
and in stock too? crap
<egg|zzz|egg>
ferram4: that's an issue, since it means I can't grab those forces to know how to integrate the vessel on my side... any idea how I might get them?
<ferram4>
egg|zzz|egg, do you handle any of the flexing, or just translation and rotation?
<egg|zzz|egg>
I let the game handle everything that happens within a pile up (connected component of the graph of vessels whose edges are collision), moving the pile up around myself
<egg|zzz|egg>
(I don't even rotate the pile up, I'd like to in the distant future but let's ignore that)
<ferram4>
So, you only need net force, no moments?
<egg|zzz|egg>
yup
<egg|zzz|egg>
hm, there are bodyLiftLocalVector things for the stock aerodynamics, so maybe I could fetch those as well as the forces
<ferram4>
Hmm... so, if you're willing to access one of my classes specifically, all you need is to iterate through each Part, get FARAeroPartModule, and then read totalWorldSpaceAeroForce from that.
<egg|zzz|egg>
ferram4: that won't work with stock
<ferram4>
I don't know how to make it work for stock.
<egg|zzz|egg>
ferram4: and I'd like to remain ferramless-compatible
<ferram4>
I avoid that as much as possible.
<egg|zzz|egg>
me too
<egg|zzz|egg>
but still
<egg|zzz|egg>
ferram4: it seems stock sets bodyLiftLocalVector, etc. which store the information about the forces that were applied to the part
<egg|zzz|egg>
ferram4: do you set those too?
<ferram4>
No.
<egg|zzz|egg>
argh.
<ferram4>
FAR and stock aero are as separate as possible.
<ferram4>
The thing is, because of the weird shenanigans that stock uses (because it won't use simple coefficients or allow any quick "set the aero force to X and do nothing else") I can't reliably set anything for stock aero.
<ferram4>
I could risk double-counting forces or something even worse.
<egg|zzz|egg>
ferram4: so bodyLiftLocalVector is only read for the aero display, I think the others are too; if you were to set it I could read that
<ferram4>
But, as a note, that's not going to be the total force.
<ferram4>
It'll just be the body lift.
<egg|zzz|egg>
hmm
<ferram4>
You'll still have wing lift and drag
<ferram4>
s/drag/wing drag and body drag
<Qboid>
ferram4 meant to say: You'll still have wing lift and wing drag and body drag
<egg|zzz|egg>
for some reason the stock lifting surfaces use part.addforceatposition, so like engines (so that would be picked up by me...)
<egg|zzz|egg>
ferram4: alternatively, could you apply your forces with part.addforceatposition, rather than rb.addforceatposition? that means I pick them up (and FlightIntegrator applies them to the RB, that's how engines etc. work)
<ferram4>
I can look into changing that. I'll have to see what the difference is in functionality and how it handles applying forces to Parts that lack rigid bodies.
<ferram4>
I know how I handle that, but I dunno how stock does.
<egg|zzz|egg>
ferram4: well, you could keep your logic but just at the last minute apply the force to the part with the rigid body instead of the rigid body (not relying on any stock logic)
<egg|zzz|egg>
bofh: apparently someone got the complexification pun :-p
<ferram4>
egg|zzz|egg, The latest dev build of FAR will support reading things through Part.Add____
<ferram4>
You must be capable of handling Part.AddForce, Part.AddTorque, and Part.AddForceAtPosition
<egg|zzz|egg>
yup, that's what I plan on doing
<ferram4>
egg|zzz|egg, Also, based on the code in RealChuteLite embedded in FAR, this will also require changes to RealChute itself. For your initial testing, do not use RealChute proper.
<egg|zzz|egg>
ferram4: for context, those were added in coordination with NK et al. because I needed a census of nonconservatives in space (but now it's nice to have it in the atmosphere too :-p)
<egg|zzz|egg>
ferram4: thanks for the warning
<ferram4>
egg|zzz|egg, Also, since this is _your_ feature request, you're gonna be the one testing it.
<egg|zzz|egg>
ferram4: other question, how do you kill off stock aero? I should probably detect if it's doing nothing so I don't pick up its forces (which come via separate channels)
<egg|zzz|egg>
ferram4: sure
<ferram4>
egg|zzz|egg, So, based on what's there, it's that I'm hijacking FlightIntegrator's UpdateAerodynamics and either not applying forces in there or I'm not calling something else that the stock version calls.
<ferram4>
There are a few things still set in stock aero as a necaessary evil to make the heating code work
<ferram4>
Don't worry about detecting stock aero. Keep checking it and when FAR is installed it'll return 0s unless it's one of the few things that has an override to allow stock aero to apply things.
<ferram4>
Like BDArmory missiles.
<ferram4>
And anything left with a stock lifting module when all is said and done just in case that's actually useful for something.
<egg|zzz|egg>
ferram4: so bodyLift etc. will be 0 in stock?
<ferram4>
Should be.
<egg|zzz|egg>
because I seem to be seeing some visualization in the debug menu when asking for visualization, but maybe you set it yourself
<ferram4>
Oh, you mean the aero arrows on the parts?
<ferram4>
Because I set those myself. All the lift is cyan, all the drag is red, and there are yellow moments going out the side, yes?
BPlayer has quit [Ping timeout: 204 seconds]
<egg|zzz|egg>
ferram4: yup
<egg|zzz|egg>
ferram4: thanks for the quick reply btw
APlayer has quit [Remote host closed the connection]
<egg|zzz|egg>
ferram4: FAR's latest build is for 1.3?
<ferram4>
Yes. Is this unacceptable?
<egg|zzz|egg>
not for development (I can start my 1.3 install), but principia continues to dual-target because RSS
<egg|zzz|egg>
I'm not sure how you've dealt with the lagging of RSS/RO on your side, haven't quite followed
<egg|zzz|egg>
(it used to be that they were about to launch on 1.3 by the time of the
<egg|zzz|egg>
чебышёв release, and now Cesàro just got out :-p
<egg|zzz|egg>
)
<egg|zzz|egg>
ferram4: tbh it's possible that by the time we're done with the 10-point plan for reunification RSS will finally be out on 1.3 and we'll stop targeting 1.2, everything takes time on the Principia side too :-p
<ferram4>
Well, I don't believe all that much needs to be updated for RO for 1.3
<ferram4>
But if necessary, I can rebuild for 1.2. It should not affect any of the code that you implement though.
<egg|zzz|egg>
yeah, I can test your current build on my 1.3 install
<egg|zzz|egg>
ferram4: I mean, 1.3 and 1.2.2 are fairly similar, we have some minimal ifdefing going on to allow for dual-building (mostly a nameWithArticle extension method hack for theName)
<egg|zzz|egg>
ferram4: yup, flight seems to be nominal above 300 m with your fix (where I put the manageability threshold for the first part of the plan for reunification), with the expected effect that I appear to be flying inside hurricane-force winds (interpreting rotating frame as inertial on Principia's end)
<egg|zzz|egg>
prior to your fix flight was parabolic past 300 m >_>
<ferram4>
Excellent.
<egg|zzz|egg>
ferram4: the next part of the 10-point plan is to correctly interpret the rotating reference frame; at that point things will work except landing above 300 m will result in Principia mercilessly driving you into the ground along your geodesic
<egg|zzz|egg>
so following that there are a bunch of steps to have ground interaction detection (and transitive ground interaction detection through collisions)
<ferram4>
Or you can just cause being on the ground above 300m to be deadly.
<ferram4>
There's no reason to need to be able to launch from whichever of China's sites is 1km up, right? :P
<egg|zzz|egg>
:D
<egg|zzz|egg>
ferram4: that would also cause moon landings to be deadly though, which is the main bug we're trying to fix with this :D
<UmbralRaptor>
So with the most realistic mods, you need to fake the moon landings? ?