2 ms·

> new standards that didn't exist when the work was done Being correct is not a new standard. > This is simply not how things are done, in science or in socie

by wyager 10y ago

> new standards that didn't exist when the work was done Being correct is not a new standard. > This is simply not how things are done, in science or in society. Oh, I say! Positively indecent! A moral outrage! We can't have our morals compromised by scientific objectivity! > we do not and will never revisit all prior work and formally re-judge whenever standards or policies change. Have you ever heard of Principia Mathematica? > nothing can be done about the old papers Except to try and figure out when they're wrong, as this bot is doing.

dahart 10y ago
Please try to judge comments in their context and not look for excuses to try and humiliate someone with sarcasm. I was discussing standards with @int_19h, who brought up the issue of new methodologies. Of course being correct isn't a new standard, I agree with you, but that's not what was being discussed. Of course morals shouldn't be compromised by scientific objectivity. Again, you're arguing a straw man - that's not the issue I was talking about. I have stated multiple times, including my first reply to you, that I think the bot is fine. My argument in context is that a paper cannot change because it has been published. Do you disagree with that? That doesn't have any bearing on whether bots or people find & publish errors later. It does have a bearing on how people will respond to PR campaigns to publish errors when nothing can be done about it on the part of the author. Statcheck will do good things for authors who get to use it before they publish rather than after. Maybe you're not reading all of what I wrote? Maybe I hurt your feelings?
- Cogito 10y ago
  My argument in context is that a paper cannot change because it has been published. I think this statement is, if not completely untrue, grossly misrepresenting how existing papers are interacted with. First of all papers, as with all publications, have errata published all the time. These errata may be included in future prints, or published in a separate location that can be looked up by people using the paper. Publishing errata is not a new occurrence, and although perhaps technically the original paper remains published unchanged, it is disingenuous to claim that this means the paper cannot change. Modern publishing methods, such as the arXiv, allow for new versions of the paper to be uploaded, literally changing the published version of the paper. As you point out yourself, literature reviews should point out issues with existing papers. Do you think that the original authors throw their hands in the air, thinking to themselves "oh well, it's published, nothing can be done"?? Of course not! If they are still engaged with the subject they either defend the paper, correct obvious mistakes, or continue experimentation or investigation in response. To claim that errors should not be pointed out simply because the original authors can do nothing about the errors is diversionary at best. Of course errors in published results should be made public. How else can we trust any of the works? If errors in existing research is always hidden, squelched, swept under the rug, we have no reason to trust it. It is the openness of research - publishing in the open, criticising in the open, discussing in the open - that allows us to trust research in the first place. Indeed, that trust is already eroded by revelations of systemic issues like p-hacking within published research. You may be suggesting that posting these analyses to the individual papers was the wrong way to do it, that it would be better done in a literature review or paper. I completely disagree. It is essential that anyone looking to reference a paper with a glaring mistake in it (which many of those affected are) is able to see that mistake and correct for it. Leaving the old research be is just ensuring that incorrect ideas are allowed to propagate, and have more of an impact than they ever should.

by wyager 10y ago

dahart 10y ago

Please try to judge comments in their context and not look for excuses to try and humiliate someone with sarcasm. I was discussing standards with @int_19h, who brought up the issue of new methodologies. Of course being correct isn't a new standard, I agree with you, but that's not what was being discussed. Of course morals shouldn't be compromised by scientific objectivity. Again, you're arguing a straw man - that's not the issue I was talking about. I have stated multiple times, including my first reply to you, that I think the bot is fine. My argument in context is that a paper cannot change because it has been published. Do you disagree with that? That doesn't have any bearing on whether bots or people find & publish errors later. It does have a bearing on how people will respond to PR campaigns to publish errors when nothing can be done about it on the part of the author. Statcheck will do good things for authors who get to use it before they publish rather than after. Maybe you're not reading all of what I wrote? Maybe I hurt your feelings?

Cogito 10y ago
My argument in context is that a paper cannot change because it has been published. I think this statement is, if not completely untrue, grossly misrepresenting how existing papers are interacted with. First of all papers, as with all publications, have errata published all the time. These errata may be included in future prints, or published in a separate location that can be looked up by people using the paper. Publishing errata is not a new occurrence, and although perhaps technically the original paper remains published unchanged, it is disingenuous to claim that this means the paper cannot change. Modern publishing methods, such as the arXiv, allow for new versions of the paper to be uploaded, literally changing the published version of the paper. As you point out yourself, literature reviews should point out issues with existing papers. Do you think that the original authors throw their hands in the air, thinking to themselves "oh well, it's published, nothing can be done"?? Of course not! If they are still engaged with the subject they either defend the paper, correct obvious mistakes, or continue experimentation or investigation in response. To claim that errors should not be pointed out simply because the original authors can do nothing about the errors is diversionary at best. Of course errors in published results should be made public. How else can we trust any of the works? If errors in existing research is always hidden, squelched, swept under the rug, we have no reason to trust it. It is the openness of research - publishing in the open, criticising in the open, discussing in the open - that allows us to trust research in the first place. Indeed, that trust is already eroded by revelations of systemic issues like p-hacking within published research. You may be suggesting that posting these analyses to the individual papers was the wrong way to do it, that it would be better done in a literature review or paper. I completely disagree. It is essential that anyone looking to reference a paper with a glaring mistake in it (which many of those affected are) is able to see that mistake and correct for it. Leaving the old research be is just ensuring that incorrect ideas are allowed to propagate, and have more of an impact than they ever should.