I was speaking of the angles in the shape of the planes design.
Have you ever seen a diffraction pattern caused by any discontinuities?
I have seen thousands of them ... the fourier transform of any discrete functions makes very nice detectable pattern that says : "NOT NATURAL". And that what radars are trained for spotting.
And the more symmetries the more the pattern reinforce itself saying: human made technology spotted. (Curie principle : symmetry in causes get propagated in the effects)
Have you seen a lot of natural object that are both discontinuous and symmetric and the size of a plane?
EDIT : I forgot to explain the relationship between convolution of signals and FT but I guess everyone is educated enough to know that convolution of signal is the base of radar detection and that Fourier transform are related to convolution.
Oh.... I thought you were talking about curves in the routes programmed into the flight computer!
AFAIK, stealth planes use all flat surfaces and no curves because a curved surface is more likely to reflect a signal from an active radar beam back in the direction of the detector instead of somewhere else.
If you shine a flashlight beam on a reflective sphere, you will always see a tiny spot of reflected light somewhere on the surface. If you shine a light beam on a perfectly flat surface, and it is not angled correctly, you will not see the reflection.
So stealth surfaces are designed to be as black (to radar frequencies) as possible, and as flat as possible. And if possible, you should make as many of your flat surfaces parallel to one another as you can manage, because each one sends a reflection of a radar beam to a new spot, and you don't want the spots to be widely dispersed.
In order to classify a radar reflection as natural or artificial, you have to detect it first.
Detection is always a business decision between true/false positive/negative, the costs of sensitivity that is non linear per stage (breakthrough are non predictable) and the costs of stealth.
And since science have improved at increasing detection by using distributed measures coupled with active correction the detection is getting easier and cheaper.
There is a pace at which technology are cheap and efficient. And money (especially when GPB is not growing) is really the nerve of war. Technologies should be used opportunistically according to their costs that can change faster than the design/delivery process.
The blindness on the cost efficiency and long term bets is taxing a lot the normal economy that in turns slow down and has difficulties feeding back military with useful cheap innovative realiable real innovations. We cannot predict the future. Especially when it comes to innovation.
And look: we rely on the best of our technology to build weapons, but these technologies don't seem to deliver.
Maybe the F35 story is an allegory of how our economies are in bad shape. And maybe the greatest military decision would be to diminish the defense/NASA budgets and fix the market and society.
PS I insist that these software problems are not unique to USA. Rafale, A400M, Typhoon, probably russian planes have been suffering "outstanding delays" & "budget slip" and software are often involved.
I do not think this is only a military problem. Something is rotten in the R&D department. Maybe the way we organize ourselves.
Have you ever seen a diffraction pattern caused by any discontinuities?
I have seen thousands of them ... the fourier transform of any discrete functions makes very nice detectable pattern that says : "NOT NATURAL". And that what radars are trained for spotting.
And the more symmetries the more the pattern reinforce itself saying: human made technology spotted. (Curie principle : symmetry in causes get propagated in the effects)
Have you seen a lot of natural object that are both discontinuous and symmetric and the size of a plane?
EDIT : I forgot to explain the relationship between convolution of signals and FT but I guess everyone is educated enough to know that convolution of signal is the base of radar detection and that Fourier transform are related to convolution.