robryk @robryk@qoto.org
I enjoy things around information theory (and data compression), complexity theory (and cryptography), read hard scifi, currently work on weird ML (we'll see how it goes), am somewhat literal minded and have approximate knowledge of random things. I like when statements have truth values, and when things can be described simply (which is not exactly the same as shortly) and yet have interesting properties.
I live in the largest city of Switzerland (and yet have cow and sheep pastures and a swimmable lake within a few hundred meters of my place :)). I speak Polish, English, German, and can understand simple Swiss German and French.
If in doubt, please err on the side of being direct with me. I very much appreciate when people tell me that I'm being inaccurate. I think that satisfying people's curiosity is the most important thing I could be doing (and usually enjoy doing it). I am normally terse in my writing and would appreciate requests to verbosify.
I appreciate it if my grammar or style is corrected (in any of the languages I use here).
Joined Nov 2020
No, absolute value of the mean rather than mean of the absolute value.
Don't you mean "pointwise multiply" when you say convolve? (Then, the _resulting signal_ would average out to 0 if the sine was absent and to a complex number with absolute value prop. to sine's strength in the original signal and arg indicating the phase offset. Note that this is about the average of the signal and not the absolute value of the signal.)
Why do you want RMS of a complex signal? What is the physical thing you are trying to model?
ISTM that the transformation that #rust does to bodies of async functions to split them into pieces-between-await-calls requires unsafe blocks (if we hold a ref from one block to another, the ref remains valid only by virtue of !Unpin around its target and so we start relying on things that cannot be expressed in the type/lifetime system for safety).
Is there a macro library/something that would allow me to do something similar _without writing unsafe myself_?
Yes, magnitude(mean) != mean(magnitude).
Mean of complex numbers is useful or not in a very contextual way :) (e.g. see why the Fourier basis is linearly independent).
It seems that numpy is slightly silly and your options for computing squared norm is either np.real(x)**2+np.imag(x)**2 or np.abs(x)**2.
Isn't `x**2` literally the square, which will be a complex value that just rotates twice as quickly?
Do you need a dome, or just a room of any convex shape with high enough ceiling? Parallax shift rates (i.e. angle change per head position linear change) are inversely proportional to distance, so the shifts should be continuous but not necessarily smooth for a non-dome ceiling. I'm not sure what's the threshold for noticeability.
If you haven't played much any, Dreamhold is a good intro.
ifdb.org is a great resource to look for games
I was thinking of giving someone a view of sky with clouds with a very large (at least tens of meters) effective intereye distance, thus effectively scaling distances down.
Technorama in Winterthur has a "distance magnifier" on the roof, which is basically a large pair of binoculars with effective intereye distance of ~2m that point horizontally and can be rotated around the vertical axis. That causes surrounding buildings to appear much less flat, so the obvious idea is doing the same to the view of clouds. I would guess that having 3 such cameras would already allow for some nonterrible interpolation for viewing angles where the line joining the eyes doesn't align with the line joining any two cameras (I suspect that being able to provide roughly correct view for slight angular movements is important for creating a realistic-looking view for humans).
Somewhat unrelatedly, have you thought about setting up 1-2 additional cameras for parallax to the clouds?
The abstract made me think that one can assign "worst-case(?) number of CAS per MCAS" to any implementation of MCAS and that they prove that: (a) lower bound for that is k (b) their implementation achieves k+1. That would be a near-optimality claim about their implementation. However, that's not the case, because there are actually two different such values: the one the lower bound (in the Impossibility section) talks about is not the one they show is equal to k+1 in their implementation.
In fact, one can probably show that any such implementation, for some adversarial scheduling and operation sequences, might need to do unboundedly many CASes for some MCAS operations (otherwise it would be waitfree).
But it does make you install semi-permanent per-operation data (the descriptors cannot be cleaned up until way later -- if you wanted to clean them up immediately you'd be back up to 2k CASes -- and even if they could you'd need to let them stay at least dereferencable-but-with-arbitrary-content for nearly as long).
Do you understand what is the thing they're providing bounds on? The lower bound seems to be on number of CASes needed pessimistically under some interleaving of contended MCASes and the upper seems to be on number of CASes needed for _uncontended_ MCAS operations.
Scratch that, I'm reading too inattentively. Still seems slightly fishy in different ways, but I'm probably still reading too inattentively.
You mean MCAS implementation where two operations on disjoint memory locations also have disjoint set of memory locations they access? My intuition is that it would be completely impossible to do waitfreely (via some similar argument to the one for Omega(thread_count) memory bound for nontrivial waitfree data structures).
Hm~ why does the lower bound assume that critical steps are CASes and not simple atomic writes? That's probably true in the setup where the number of threads is not bounded, but otherwise I see no obvious reason for that to be true.
@koakuma Sadly only lockfree and not waitfree.
I don't mean the CG shift impacting how the thing flies, but impacting whether the recoil only pushes the drone backward, or whether it also rotates it. Do you expect recoil-generated torques to be compensatable with dynamic control without rotating by more than a few degrees? (Seeing that it fires in bursts.)
@kravietz Ah, maybe it's shooting blanks? (I don't really know how the "blank adapter" should look like.)
@kravietz I'm surprised that the center of mass change due to getting rid of rounds doesn't cause the drone to be slightly rotated by the gun either during first or last shots.
I enjoy things around information theory (and data compression), complexity theory (and cryptography), read hard scifi, currently work on weird ML (we'll see how it goes), am somewhat literal minded and have approximate knowledge of random things. I like when statements have truth values, and when things can be described simply (which is not exactly the same as shortly) and yet have interesting properties.
I live in the largest city of Switzerland (and yet have cow and sheep pastures and a swimmable lake within a few hundred meters of my place :)). I speak Polish, English, German, and can understand simple Swiss German and French.
If in doubt, please err on the side of being direct with me. I very much appreciate when people tell me that I'm being inaccurate. I think that satisfying people's curiosity is the most important thing I could be doing (and usually enjoy doing it). I am normally terse in my writing and would appreciate requests to verbosify.
I appreciate it if my grammar or style is corrected (in any of the languages I use here).
Joined Nov 2020
