Slashdot Mirror
IBM Slows the Speed of Light
dptalia writes "According to an article on ZDNet, IBM has come up with a way to slow light to 1/300 of its normal speed. While this has been done in laboratories before, IBM has found out how to do this using standard materials, which opens the possibility of mass production. This means that the dream of having optical based CPUs may be closer than previously thought." From the article: "When the optical conversion might start to occur is a matter of speculation. Luxtera has said it will start to commercially produce products in 2007. The computer industry, however, tends to move slowly when it comes to major overhauls of computer architecture. Several components will have to be developed before photons can replace electrons inside computers. A paper providing details on the chip will run in Nature on Wednesday."
That hardly proves that it can't be done; people used to see no way that a plane could possibly go faster than sound.
Don't blame me; I'm never given mod points.
The wavelength of an electron is extremely tiny compared to the wavelength of light. This means that feature sizes for light based chips are necessarily much larger than those for electron based chips. Barring some advancement that allows us to pack more functionality per unit area into an optical chip, optical computing will remain a very niche field.
The best way to accelerate a windows box is at 9.8 meters per second square.
Actually, the sound-barrier analogy is misleading. For the speed of sound, people KNEW that things could exceed that speed long before we got planes to do it. The issue was one of technology: could we build a plane to withstand the stress?
For the speed of light issue, it's a different. If you believe Einstein's Theory of Special Relativity, you just can't exceed that speed. At least not if you start below light speed and remain in this universe. There's a very clear physical law that prohibits this, not a concern about technology being up to the task.
Of course, the law might be wrong. Or there may be ways of side-stepping it. In fact, I'm giving a whole planetarium talk this very evening on that very issue.
In the same way that Jesse Owens with a twisted ankle is faster than Fat Albert.
which is science. Blind postulates are not.
Most of the time the light would still be operating on the good old light speed. But for it to work in computers you still would have to slow it down in places and even stop it. For example to let another beam of light to pass before it can go through.
I don't quite see where you're getting this idea from. It's a bit barmy to imagine current electonic processors firing two lines at the same time and then having an 'electron traffic light' to let one signal pass by making another wait. This may sound like a switch but it's not, because at no point are you actually 'stopping' electrons. If you don't produce a voltage your electrons aren't going to move in a current, so you haven't stopped them because you never fired them to start with. As i'm sure you realise, in digital electronics data transmission is acheived by voltage state. Changing state from 0 to 1 happens because you apply a voltage, and 0 to 1 because you stop applying it. With photonics, the equivalent must be turning the source on and off?
It may be beyond my knowledge of physic's but slowing down light within an optical processor (to better interface with other devices or traditional electronics or whatever) sounds like an alternative to having light signals running at a lower frequency (more time spent in each state so peripherals can spot signals). Slowing down light and introducing a delta velocity surely means we need a way to buffer light, much like a capacitor stores charge?