3 ms·

Well, one way or another, once you get reproducible experimental results that conflict with known physics, someone has to do the due diligence in finding the fl

Well, one way or another, once you get reproducible experimental results that conflict with known physics, someone has to do the due diligence in finding the flaw, it would be irresponsible to shrug and say "ah, must be a fluke, who cares why". Perhaps if we had a whole range of other promising avenues to explore for new phsyics (like natural scientists a century or two ago), then it would be different, but as of now it seems there's not much competition that's more promising than the EmDrive was.

jfengel 6y ago
That's the ideal, but we actually do it all the time. No real experiment actually corresponds precisely to the model. Experiments go Just Plain Wrong all the time, and we don't look at them; we just assume we made a mistake in the setup and repeat it. If we engaged in due diligence for every single thing we noticed, we'd do nothing else. Science is much messier than the neat models we set ourselves. There's a whole subdomain of studying the Demarcation Problem, which is trying to figure out the difference between "hypothesis revision" (science) and "goalpost moving" (pseudoscience). Turns out it's harder to define rigorously than you'd expect it to be. A huge number of scientific decisions end up being made on a combination of "who has money to spend", "who has time", and "who is interested". That's utterly non-rigorous, but it's actually pretty fundamental to how real science gets done. I'd say there's a lot of stuff more promising than EmDrive. We all knew it was a mistake, and if it hadn't somehow attracted some prominent backers and a ton of people hoping that this time we were gonna get green Orion slave girls in metal bikinis, it would have been ignored years ago. We could have spent that money and time on other things, but they would have been boring even if they'd worked out.
- PeterisP 6y ago
  "I'd say there's a lot of stuff more promising [...] but they would have been boring even if they'd worked out." - so which is it? That's kind of my point, we have a bunch of ideas that are more likely to succeed, but anything that would be boring even if they'd worked out aren't really promising in the "can it result in a meaningful impact" sense; and there seems to be a lack of experimental ideas that are within our grasp and could potentially reveal something interesting instead of just re-confirming thins that we're almost certain about.

jfengel 6y ago

That's the ideal, but we actually do it all the time. No real experiment actually corresponds precisely to the model. Experiments go Just Plain Wrong all the time, and we don't look at them; we just assume we made a mistake in the setup and repeat it. If we engaged in due diligence for every single thing we noticed, we'd do nothing else. Science is much messier than the neat models we set ourselves. There's a whole subdomain of studying the Demarcation Problem, which is trying to figure out the difference between "hypothesis revision" (science) and "goalpost moving" (pseudoscience). Turns out it's harder to define rigorously than you'd expect it to be. A huge number of scientific decisions end up being made on a combination of "who has money to spend", "who has time", and "who is interested". That's utterly non-rigorous, but it's actually pretty fundamental to how real science gets done. I'd say there's a lot of stuff more promising than EmDrive. We all knew it was a mistake, and if it hadn't somehow attracted some prominent backers and a ton of people hoping that this time we were gonna get green Orion slave girls in metal bikinis, it would have been ignored years ago. We could have spent that money and time on other things, but they would have been boring even if they'd worked out.

PeterisP 6y ago
"I'd say there's a lot of stuff more promising [...] but they would have been boring even if they'd worked out." - so which is it? That's kind of my point, we have a bunch of ideas that are more likely to succeed, but anything that would be boring even if they'd worked out aren't really promising in the "can it result in a meaningful impact" sense; and there seems to be a lack of experimental ideas that are within our grasp and could potentially reveal something interesting instead of just re-confirming thins that we're almost certain about.