We, the community, will applaud your growth, celebrate your new strength, and leave your mistake in the past where it belongs.
Does this mean we should past errors, as long as they’ve been admitted? As I understand it, there is a chain of three different claims:
(1) LWers should admit their mistakes.
(2) To encourage this, LWers should be supportive of those who admit their mistakes.
(3) As a mechanism for this, LWers should ignore errors that have been admitted as mistakes by their authors.
I like (1) and (2). I’d like to spell out the cost of (3).
Past mistakes are relevant to estimates of future accuracy because patterns in mistakes could indicate bias or ignorance. The author’s admission of error should not change this (though it should make us discount her other output less, because her demonstrated willingness to admit errors decreases the likelihood that her other output has errors known to her but not yet to us).
Thus, not accounting for past mistakes (whether others’ or our own) prevents the discovery of, and thus correction for, error patterns. Hence the cost of such an enforcement mechanism.
That being said, the costs might be outweighed by the benefits, if this mechanism is much better than other mechanisms.
On a related note, people do occasionally try to keep track of the errors of public figures, such as at TakeOnIt (previously mentioned on LW here and Project Votesmart.
Ok, if someone constantly has mistakes to admit, that would be a problem, and yes, we should encourage that person to be more careful in forming their positions.
Also, though I endorse the typical behavior, what I said about how LW responds to mistakes is a description of what currently actually happens, and my message is directed at those who are afraid the response might be harsher.
Does this mean we should past errors, as long as they’ve been admitted? As I understand it, there is a chain of three different claims:
(1) LWers should admit their mistakes. (2) To encourage this, LWers should be supportive of those who admit their mistakes. (3) As a mechanism for this, LWers should ignore errors that have been admitted as mistakes by their authors.
I like (1) and (2). I’d like to spell out the cost of (3).
Past mistakes are relevant to estimates of future accuracy because patterns in mistakes could indicate bias or ignorance. The author’s admission of error should not change this (though it should make us discount her other output less, because her demonstrated willingness to admit errors decreases the likelihood that her other output has errors known to her but not yet to us).
Thus, not accounting for past mistakes (whether others’ or our own) prevents the discovery of, and thus correction for, error patterns. Hence the cost of such an enforcement mechanism.
That being said, the costs might be outweighed by the benefits, if this mechanism is much better than other mechanisms.
On a related note, people do occasionally try to keep track of the errors of public figures, such as at TakeOnIt (previously mentioned on LW here and Project Votesmart.
Ok, if someone constantly has mistakes to admit, that would be a problem, and yes, we should encourage that person to be more careful in forming their positions.
Also, though I endorse the typical behavior, what I said about how LW responds to mistakes is a description of what currently actually happens, and my message is directed at those who are afraid the response might be harsher.