I'm surprised (but not really) at the amount of resistance to to questioning these results, when at the very least you have to agree with the following:
This method AND the prior method can't BOTH be right.
That means at least one of them is WRONG. Had I questioned the prior method's claims and techniques yesterday I would have had the same response from the same people.
Science (to some) is still about questioning everything. It scares me how many people on HN have lost sight of that.
Your comment didn't read like an ode on the scientific method to me. Instead, I saw a cheap dismissal of the research in it. And the article even addresses your concerns:
> "They're using an indirect method," he said. "We're using a direct measurement and we're measuring from a mineral that has actually trapped atmosphere. It's a huge step forward, really."
The previous number was an estimate derived from other, indirect findings. In other words, "given what we know, 2.x% O₂ seems most likely and plausible" versus "we dug up some salt with air from the time trapped in it, and measured 10.9% O₂". So they did exactly what you're referring to: questioning previous results and giving evidence to support their claims. And yet you question their results without providing any evidence. Your criticism boils down to "well this isn't what we used to believe". How else could science progress if not by surpassing what was previously known or accepted as the best answer to date?
Science is about questioning stuff, coming up with new ideas and backing them up with evidence, not about making dismissive comments on the internet.
Wrong. See, this is the type of thing I'm talking about. There is only one way to make a direct measurement and it involves a time machine.
I'm not saying that to be pedantic. That is only way to prove this technique and/or others like it are absolutely correct. While I applaud this man's effort and his approach, I'm skeptical of the results given the difficulty of problem at hand and methodology used.
Let's make sure to call it what it is. Another method of estimation.
We can never reach absolute certainty. Improving existing approximations of the truth is the best we can hope to achieve. If you only accept absolute unquestionable truth then I recommend you have a look at the history of science. Lots of what was the state of research in the past is now known not to have been entirely correct, sometimes even outright false. But we need to make incremental progress towards the goal of understanding the world, even if we sometimes get it wrong. That is the moly thing that opens the path for new discoveries. Following your reasoning, these surpassed results are worthless and were no good even when they were found. But in the grand scheme of things they are what enabled later researchers to uncover better and more accurate results.
Oh and if you weren't so focused on absolute truth, you might have noticed that natural sciences concern themselves with models of reality, because nature is freakishly complex and if one doesn't settle for some approximation of reality. Otherwise one wouldn't get anywhere trying to solve the simplest problem.
Models in natural since need to be cross verified by multiple findings using various methodologies. Each model and method adding to the evidence of our understanding and the probability that we got it right.
In this case, the new methodology refutes the prior one. So, instead of adding the evidence that current models are right, this creates a new hypothesis, for which currently only this one model seems to agree with.
So, I get what you're saying about there will never be an absolute truth, but I hope you get what I'm saying about how this one finding alone means little in terms of being conclusive.
Being right for the wrong reasons doesn't win you any points.
If you had questioned the prior method's claims and techniques yesterday based on this same sort of "what about this obvious problem?" then you would have had the same response and deservedly so.
If I declare that the Earth is round because I drew a circle on a piece of paper, I can't wave away criticism of my method just because the result is correct.
All extrapolations of this type are difficult to prove out. They are based on a lot of assumptions. Those assumptions may not all be right and some critical components might be missing. The model being used to compute slippage over time might be overfit to the parameters.
If I declared the world to be flat then proved it by going to a truly flat field to provide parameters to a mathematical model that showed I was correct, it wouldn't make me correct no matter how convincing my credentials or my paper.
You have no idea if you're right or if this conclusion is actually correct. If the conclusion is wrong, you have no idea if it's wrong because of the problem you posted, or if it's wrong because of something completely different.
Virtually all science is wrong in some respect. You can get a great batting average by just saying "no, wrong" to everything you see. That doesn't make it useful, nor does it make your reasons for saying "wrong" correct.
This method AND the prior method can't BOTH be right.
That means at least one of them is WRONG. Had I questioned the prior method's claims and techniques yesterday I would have had the same response from the same people.
Science (to some) is still about questioning everything. It scares me how many people on HN have lost sight of that.