Take three light valves. They need to be very fast. Put them in a stack. Sequence them so that they open and close one after each other, timed such that each one opens and closes a certain amount of time after the previous one, where that amount of time is the time it takes light to travel from one valve to the next.
This is analogous to the three-solenoid peristaltic pump for fluids described in [1], except that it doesn't pump the light along—it just allows it through in one direction but not the other, by timing the valves to match the propagation of the light. It does, however, result in a reduction in effective brightness of the light it allows through in that direction, by only allowing 1/3 of it through, and produces a very fast (invisible to humans) flickering.
49/343 [2018-05-17]-- notexactly, May 19 2018 [1] Three-solenoid peristaltic pump [PDF] https://abcm.org.br...ORS/SSM4_VII_07.pdf [notexactly, May 19 2018] One-way Glass The idea that inspired this one. [notexactly, May 19 2018] scissors that are not faster than light http://math.ucr.edu...ty/SR/scissors.html [beanangel, May 19 2018] How fast? How close together?-- pocmloc, May 19 2018 One light-nanosecond is about one foot. A rapatronic camera's shutter can open for as short a time as 10 ns according to Wikipedia. I don't know at the moment how the three valves' timings will affect the math, but this gives an order-of-magnitude estimate of ten or twenty feet. Of course, faster light valves might have been developed since then.-- notexactly, May 19 2018 This is quite a cool idea, at least as a thought experiment.
[+]-- MaxwellBuchanan, May 19 2018 Well... that's conceivably nuts.
I'ma posit that given perfect reflectors, light will still make its way out (though even then you'd be talking a 20ft input path and (at least) a 40 ft exit path which is still really weird)-- FlyingToaster, May 19 2018 A Kerr cell shutter can open and close in nanoseconds, and has the secondary benefit of involving interesting chemicals with "nitro" in their names.-- Wrongfellow, May 19 2018 Neat. [+]
One issue is that the timings will be different for off-axis light. If the timings are too tight, you would only be able to see a single spot of light through the window.-- mitxela, May 19 2018 // One issue is that the timings will be different for off-axis light. //
That's why I didn't put it in home:window:privacy like the inspiring idea. I thought this would be more useful in scientific and engineering applications, where you can guarantee on-axis light.-- notexactly, May 19 2018 How is this different (in its effects) from a single light valve that's open one third of the time?-- pertinax, May 19 2018 How do you synchronise their opening and shutting? Seeing as how there will be a time lag from the open and shut signal travelling down the wire? Do you need each one connected to its own atomic clock? Or can you offset the synchronisation signal to accomodate the predicted time delay? Or could the pulse of light itself be used to synchronise in one direction?-- pocmloc, May 19 2018 It's not beyond the wit of man to model the delay of pulses propagating down transmission lines.
So, you basically just cut a few bits of coax to the correct lengths, and connect them all to the same pulse generator.
Obviously, the coax length needs to be a bit longer than the speed-of- light distance from the pulse generator to the shutters!-- Wrongfellow, May 19 2018 I am reminded of the faster than light big scissors, the tips of which travel faster than light [link] Just make one, or three, of these and see what it does. Oh, oops, the scissors thing has been refuted...
There is another scheme that might be called marquee lights...-- beanangel, May 19 2018 It's not the tips that travel faster than light (if they approached the speed of light, they'd become infinitely massive, and you couldn't move them). It's the contact point between the blades that travels faster than light. And the contact point can indeed travel at any speed, including faster than light.-- MaxwellBuchanan, May 19 2018 Interesting idea. There are a few simple refinements that might push this towards practical usefulness.
First, you don't need 3 shutters, just 2, and you can make it work with a theoretical best case attenuation of only 50%.
Assuming a perfect 10ns shutter open time, have shutter 1 and shutter 2 separated by 5 light nanoseconds. open shutter 1 from time 0 to time 10ns. Open shutter 2 open from time 5ns to time 15ns. Repeat the cycle every 20ns. Light passing through shutter 1 from 0 - 10ns, will arrive during the shutter 2 open time of 5 - 15ns. Light passing through shutter 2 at 5ns to 15ns will arrive at shutter 1 when it is closed between 10ns and 20ns.
Another optimization is to put a material with high refractive index between the shutters. This would have two benefits. First, the light will travel slower, requiring less distance. Second, the refraction would somewhat straighten off-axis light. Using polycarbonate with a refractive index of 1.6, slows light down to 188,000 km/s, reducing the 5 light ns distance from 1.5m to 0.94m. Depending on your budget, you might try a more exotic materiel with an even higher index.
If 1 meter is still too thick, you can't find any better shutters, and you're willing to have more attenuation, you can play with the shutter timing to make it work. For example, say we can have only 2 light ns between the shutters. In that case, if shutter 1 is open between 0 and 10ns, and shutter 2 is open from 8ns to 18ns (still repeating every 20ns), light will get through for 4 ns every 20ns, so there will be 80% attenuation.
Off axis light isn't too big of an issue. Let's say we want to deal with light at up to a 45 degree angle not using a refractive material. at a 45 degree angle, it will take 7.07ns instead of 5ns to go between shutters. In order to prevent light going backwards we'd need to increase the time when both shutters are closed by 2.07ns (from 20ns to 22.07ns). This will of course slightly increase the attenuation of light passing in the desired direction. For the first timing example 10/22.07 = 45% of on axis light will be transmitted. Off axis light at a 45 degree angle will have (10-2.07)/22.07 = 36% transmitted.-- scad mientist, May 20 2018 Hmm, split zer old photon, send one bit down the old Peristaltic one-way light valve and see what happens.
//scissors that are not faster than light
Well if they were, we wouldn't be able to see 'em. Tieing bits of string on them would aid retention.-- not_morrison_rm, May 20 2018 Are there relativistic issues to do with synchronising time to this accurately in different places?-- pocmloc, May 20 2018 Thank you, [scad mientist], for doing a bunch of math in support of the idea. I didn't feel up for doing it. Also, I had thought of the two- shutter version, but I didn't feel like doing even the mental simulation to decide whether it was viable, and I forgot to include any mention of it in the idea. (I'm just glad I remembered as much as I did; I forgot most of the points I wanted to include right before I typed it up, and had to struggle to remember them.) So thanks for bringing that back up and showing its viability. I think all of that qualifies you to be a co- inventor on the patent (lol).
// Are there relativistic issues to do with synchronising time to this accurately in different places? //
No. Well, you might have to consider relativity in designing the system, but it would not be a fundamental obstacle to the possibility of doing so.-- notexactly, May 25 2018 Could you do this with a simple mechanical device?
Imagine a thick disc of opaque, light-absorbing material, fixed on an axle. Now drill holes through the disc. The holes are angled, relative to the direction of rotation. Now we shine light at just one part of the disc (say, a sector from "2 o'clock" to "4 o'clock").
If the disc is stationary, and light hits the disc face-on, no light will pass through because it will hit the (angled) walls of the holes. However, if the disc is spinning at the right speed, light *will* pass through the holes in one direction, because the angled hole will be constantly moving into alignment with the photon as it passes through. But light should not pass in the other direction, because the angled hole will be moving out of alignment.
The only problem will be that the disc has to be very thick, or it has to spin very very fast, or has to have very narrow holes (probably limited by the wavelength of light, otherwise diffraction and stuff will kick in).-- MaxwellBuchanan, May 25 2018 That's analogous to how the speed of light was measured by a few historical scientists, so I expect it'd work.-- notexactly, May 25 2018 //Imagine a thick disc of opaque, light-absorbing material, fixed on an axle. Now drill holes through the disc.//
I immediately thought of the dual discs in a spinning disc confocal microscope, or rather the two-way path through the lower Nipkow disc. I remember a conversation with a chap from Zeiss who said they had to take into account the delay in the emission light relative to the excitation. Pretty snazzy at the ~10-20 cm light path distances they deal with. However, I expect MOST of that delay is the kinetics of fluorescence emission lifetime I expect. We have lasers that will happily control light pulses down to femtoseconds, dividing by 10^-15 gives satisfying power numbers.-- bs0u0155, May 25 2018 //MOST of that delay is the kinetics of fluorescence emission lifetime I expect. // Fluorescence lifetimes are typically oto a nanosecond, which is one light-foot. So that's significant but not huge compared to the light path.-- MaxwellBuchanan, May 25 2018 random, halfbakery