You know one answer to your problem. There are others. The state you put the system into describes the solution to your problem, You let the system evolve in time, and, assuming you don't add any energy (which would destroy your computation), the system will always describe *a* solution to your problem, but not necessarily the solution that you started out with.

Yes, but that exponential relationship is slowly becoming linear as the bulk of the $$ is becoming energy.

True. But these people forget to apply the rule to the rule itself: they see that correlation itself is correlated with lack of causation, and thus have started to believe that correlation causes a lack of causation.

Also, diffract:

At the atomic level, x-rays at the right wavelength diffract around atoms (really, at that level, it's "electron clouds"), and you can use the diffraction pattern to estimate the localized density of the electron clouds, in an attempt to figure out what atoms go where (heavier atoms have heavier electron clouds). However, hydrogen atoms (protons) also tend to form a sort of cloud, but that's more of a physics limitation then a measurement one. And yes, the sample does "cook" in the process, often quite thoroughly.

My point is, the reflective/refractive/absorptive/transmissive/diffractive behavior of light doesn't depend on the frequency of the light. It depends on the characteristics of a given material for that frequency of light. That's where you get color from, incidentally.

How are they measuring the phase, though? Have they solved the phase problem?

The first rule of computer architecture is that any problem can be solved by an additional layer of abstraction.

(The corollary to this is: ...except for too many layers of abstraction.)

Without doing research, I see two possibilities. Either:

a) The relevant patent has expired. Or
b) They have prior art.

This has been going on in one form or another for >20 years. This means that regardless of whether the patent was filed, it is either expired or invalid.

Look. We have to pay for the stuff we use. We can either pay the government, or some other entity. It doesn't matter. Paying the government for roads, etc. is generally a good thing, because people don't want to be nickel-and-dimed every time they drive down the street. Large corporations in general don't want to be nickel-and-dimed for each segment of shipment they make. Police aren't privatized and we don't have "legal insurance" because that would be annoying. Education isn't privatized, because we like the idea of equal opportunity.

But we still have to pay for it all, and that's where taxes come in. In general, someone smart noticed that people who make more money use up proportionally more of the resources: they buy more things, so they use the roads disproportionately more; they have more things stolen from them, so they use the police disproportionately more, etc. Megacorporations do the same thing to an even greater extent. This is why we have a progressive tax. A "flat tax" would be completely unfair to the poor, as they are the ones who use the least amount of resources, and have the least means to pay for what they do use.

The total income of the US is somewhere around $6.8 trillion (via wikipedia, mean household income * number of households). 10% of this is not enough at our current spending rates.

However, all this is beside the point. The point is that we're taxing the thing we want to encourage: making money and spending it. Making money and spending it is what drives the economy, and we're discouraging it by taxing it. That's a bad thing. We shouldn't have an income tax, or a sales tax, or a consumption tax, or a transaction tax (although if we want to discourage HFT, that might be a thing to look at). We should have a savings tax. Of course, saving for retirement is different, but if you're just sitting on money for no reason at all, other than your fear of losing it, you're hurting the economy. This is the behavior we want to discourage, so we should tax the hell out of it. This has the benefit of the poor who make just enough to get by not paying much, if at all, and the rich-beyond-all-belief being taxed the hell out of IF AND ONLY IF they're not doing anything useful, largely making it a progressive tax, which is a Good Thing (tm). After all, if they're not going to do something useful with their money, why let them keep it?

That said, of course there are flaws in this system that need to be fleshed out, but it's at least something worth looking into.

What's the difference? 250GB in 31 days is 783kbps. Sure, you might get higher burst speeds, but if you can't supply 6.4 terabytes of data every month, you shouldn't be allowed to advertise 20Mbps speeds.

Dude, he's a Brit. He understates things. It's supposed to be read,

Except that it's impossible to know what page will actually be loaded without actually fetching it from the server (i.e., if it tries to fetch another link on 404), and it's not possible to do that without potentially changing state on the server (not every website respects GET vs POST semantics).

Not sure if troll.

Pipelining doesn't affect latency, only throughput. And the throughput increase is reflected in the increase I'n frequency. 3GHz still means that you can only perform 3000 instructions in a us.
And that assumes none of the instructions trip an interrupt, which requires waiting while the pipeline flushes, and then continuing an interrupt handler; and none of the instructions cause a TLB miss, and none of the instructions cause a cache miss, and all instructions are effectively independent of each other (Hint: none of these situations actually do). Multiple issue and multiple functional units help with getting more out of the pipeline in the case where you have enough instruction independence and you suffer cache and TLB misses. Interrupts, however, still require you to flush the pipeline to guarantee proper operation. Now imagine what happens if every 20th instruction is an API call which triggers a system call by means of an interrupt because it requires privileged access to the GPU.
I didn't RTFA yet, but that seems to me like the gist of the problem.

And releasing the middle mouse button over a link after scrolling. That annoyed the bejeezus out of me and was the reason I switched back to Firefox. It basically made ad- and link-ridden sites unscrollable with the middle mouse button.

Silly people. I can prove it for arbitrary N! Just so long as P is zero, that is.