I think that the people who are actually doing the real work here (i.e. the scientists) all have fairly realistic expectations. The rest of us can party if it makes us feel better, and it won't hurt if the end result is increased funding for science in general. And if nothing happens in the end, well, there won't be any more articles, and in a month everyone except for those genuinely interested will forget it was even there (well, there will also be the occasional science freak posting about it on Slashdot in every future story on space propulsion, but that's what Slashdot is for).

Well, except for all these.

Yes, of course. I don't support this ridiculous notion that we shouldn't try to get something useful out of the experiment just because we can't explain the theory.

Let me guess, you're still bitter from WinME?

Look at what's actually making money. Hint: it's mostly Office, not Windows, and it has been that way for many years now. Why would Office for iOS or Android, say, make any less money than Office for Windows?

Or, say, Azure. It's a money maker, despite playing catch-up with AWS.

Yeah, that's why Code runs on Linux: to let people who use Linux make apps that only run on Windows. Makes perfect sense.

If Microsoft ends up making money doing applications for Linux, it means that Microsoft has won, as well.

The nice thing about this is that there doesn't have to be a losing party.

Are you saying that you would completely ignore the repeatedly reproducible result of an experiment if there were no good theoretical physical explanation for said result?

I mean, it's your choice, but it sounds extremely stupid. If the thing works, figuring out why it works is definitely a very interesting question well worth devoting resources to, but making it useful doesn't require fully understanding the theory.

Note that this is the fifth experiment so far that has reproduced the effect (and every new experiment tries to account for some explanation that could possibly invalidate the previous one; e.g. for the last one, they ran it in vacuum).

A few physicists have already ripped it apart (and the fact that the physics is unsound is why it took so long for anyone to actually try to properly reproduce it). Basically, the thing may well work, but if it does work, it's very unlikely that the explanation that its inventor has provided is legitimate.

Of course energy has momentum. Photons, in particular, have momentum. That's why there's nothing strange about the experiment as you describe it - you increase the momentum of your rocket, but that increase is exactly counterbalanced by the momentum of the emitted photon. And when that photon hits something eventually, it will transmit its momentum to that thing etc. Overall, momentum is conserved, not just "right now", but at any future point. This doesn't seem to be the case with this engine.

I assume the question was rather whether acceleration (i.e. force) scales linearly with power input. In the experiment, it did not.

The acceleration graph is definitely not linear. They have tried to model the relation so far, and if it holds, then it will peak at about 1 newton per watt, which is insane if it really works that way - forget ion thrusters, we could throw away the car engines!

The "intended environment" is not necessarily space. If it produces thrust (which it does, since that's what they are measuring), it doesn't matter where it does that, it's still useful work that can be tapped.

"They" are many different people here. The experiment has been reproduced by, what, five different teams all across the world by now? As I understand, only the guy with the original idea is making outlandish claims; everyone else is just trying to figure out what the hell is going on.

One observation that the other teams did make is that the observed output seems to be scaling nonlinearly with input, which implies that there's a peak of efficiency. They have a model that tries to guess what that is, which seems to be consistent with the results to date. If that model is right, the peak efficiency is very high - high enough for practical uses not just in space. That is speculation, of course, but it's strictly fact-driven (and comes with a very big "if" - if the engine actually works by itself).

The rest of it is the usual journalist pop sci.

A good thing then that one of the pieces of Atom that this replaces is the editor ...

Like GP said, it's a fork with significant changes. The editor is replaced entirely, and then there's the whole .NET integration story.