Feature proposal: Close comment as resolved
post by Viliam · 2022-04-15T17:54:06.779Z · LW · GW · 15 commentsContents
15 comments
There is an article. It contains a typo. Someone mentions it in a comment. The comment is upvoted as helpful. Author fixes the typo. Now there is a no-longer-useful comment in the discussion, sometimes with replies (the author saying "thanks, fixed", etc.). Sometimes it sits at the top of the discussion, if it was sufficiently upvoted.
If I try to avoid this, I send the author a private message. This does not clutter the discussion, but the disadvantage is that other people don't see me doing this -- so they are likely to also send private messages to the author, and maybe someone will make a comment anyway.
Downvoting the comment gets it out of the way, but it's not fair to the person who made it. We should encourage reporting typos, definitely not punish it.
*
My proposal is to make a new functionality that approximately means "this comment was useful at the moment it was made, but it has outlived is usefulness". Available in the three-dots menu in the upper-right corner of the comment.
A comment can be marked like this by the person who made it, or by the author of the article. (Or by the moderators, of course.) A comment marked like this will be moved to the bottom, regardless of the sorting mode, and will by default appear collapsed. (Maybe also displayed using a different color.)
Unlike retraction, this has no impact on voting, etc. (I am not 100% sure how retraction works.) You can still read the comment, even upvote it if you wish. It just doesn't get in the way. Also, semantically, there is a difference between "I no longer support this" and "this is no longer relevant, but it was when I wrote it".
15 comments
Comments sorted by top scores.
comment by Dustin · 2022-04-15T19:18:26.189Z · LW(p) · GW(p)
Your suggestion is way easier to implement, and I like it, but just dreaming here...
It'd be neat to have functionality like Google Docs with commenting and edit suggestions within the text and margins specifically for editor-type of stuff. The post author could set some sort of permissions on who could do edit suggestions and who could view them.
This has the benefit of keeping comments for commenting on the content and putting editing in to a common editing paradigm.
Replies from: MondSemmel↑ comment by MondSemmel · 2022-04-16T15:40:59.290Z · LW(p) · GW(p)
That's a good suggestion. Fortunately for us, this already exists [LW · GW] since a couple of weeks (albeit in beta).
comment by jimrandomh · 2022-04-15T18:31:55.935Z · LW(p) · GW(p)
This seems conceptually pretty close to the "retract" feature, which, now that you mention it, should probably also be handled in comment-sorting to move it to the bottom.
Replies from: TLWcomment by Nathan Helm-Burger (nathan-helm-burger) · 2022-04-16T17:06:43.295Z · LW(p) · GW(p)
I think this is a nice idea, especially if it's just a karma-sorting effect, and it's either author-only or implements the author-can-undo system mentioned by Viliam. But actually, I came here to say that I'm even more excited for a feature which is nearly the opposite: I want comments to be able to be tagged by the author as 'ToDo'. This would let readers know that the author has seen the comment and intends to respond, and the author could have a 'ToDo' page which listed their marked Todo flagged comments either chronologically or karmically. You could even allow readers to up/down vote the ToDo flag to help the author prioritize their responses.
comment by hath · 2022-04-15T23:53:44.029Z · LW(p) · GW(p)
I'd like to have the ability to leave Google-Doc style suggestions on normal posts about typos; seems like something that might be superior of our current system of doing it through the comments? Removing the trivial inconvenience might go a long way.
Replies from: Pattern↑ comment by Pattern · 2022-07-01T16:26:58.245Z · LW(p) · GW(p)
It's still a trivial inconvenience sometimes, but:
Two tabs:
one for the response comment writing as reading
one for the reading
Note, sometimes people downvote typo comments. Doesn't happen often, but, sometimes it seems like, when the author fixes it, it happens?
comment by Mawrak · 2022-04-15T23:31:23.749Z · LW(p) · GW(p)
For typos there should be an option to just select the error in the text and submit it to the author though the web page. That's what they do on some fanfiction websites. The only downside is that a troll could potentially abuse the system.
comment by TLW · 2022-04-15T19:31:15.559Z · LW(p) · GW(p)
Only as long as there is an option to turn all of this off.
I am strongly against information-hiding in general, and this is not an exception.
Replies from: jimrandomh, Viliam↑ comment by jimrandomh · 2022-04-15T20:40:14.165Z · LW(p) · GW(p)
There's a no-free-lunch problem with sorting comments; everything that we display pushes other things further down the page, hiding them from people who don't scroll that far (and very few people scroll all the way to the bottom). In practice, a lot of features that superficially seem like hiding things, like truncating long comments behind a Read More click, actually do the opposite, preventing things from being hidden. The tradeoffs are inescapable.
(Including the tradeoff with development time, which is why there aren't more customization options. There is an accessible API, an open GitHub repo, and a third-party client, though, if you're into that sort of thing.)
Replies from: TLW↑ comment by TLW · 2022-04-15T21:11:32.882Z · LW(p) · GW(p)
There's a no-free-lunch problem with sorting comments
Absolutely, for anyone that doesn't read all comments. Hence: option.
Human preferences are many-dimensional. Everyone is an outlier in something. Something that targets the median in all dimensions results in something that works well in all dimensions for (almost) nobody.
(Including the tradeoff with development time, which is why there aren't more customization options. There is an accessible API, an open GitHub repo, and a third-party client, though, if you're into that sort of thing.)
I've been burnt one too many times by site APIs being disabled or neutered over time by websites that formerly had open APIs to really want to head down that path.
(Also, "if you want anything build your own frontend[1]" is an interesting answer to 'why isn't LessWrong more popular[2]'. I find this site clunky in general, and I am here largely despite it. I wonder how many people[3] end up simply leaving instead.)
↑ comment by Viliam · 2022-04-15T22:21:55.078Z · LW(p) · GW(p)
Which part do you object to? Collapsing threads? Sorting? Both?
EDIT:
There is already a user setting "Do not collapse comments to single line", it makes sense that this would also apply to editorial comments.
You can sort comments by karma or chronologically... it would probably make sense to keep editorial comments at their original place when sorting chronologically, and moving them to the bottom when sorting by karma (as if applying a huge negative pseudo-karma to them).
Would such implementation address your concerns?
Replies from: TLW↑ comment by TLW · 2022-04-16T11:08:05.854Z · LW(p) · GW(p)
Which part do you object to? Collapsing threads? Sorting? Both?
Both, although collapsing more than sorting. A few things -
- I would have far less objections if this was strictly about typos; it isn't[1].
- I would have far less objections if the comment author was the only person that could mark a comment as such. Unfortunately, I recognize that in the good-faith case this somewhat reduces its utility.
- It's a way for a post author to unilaterally make responses 'go away' in a somewhat less blatant manner[2]. As long as the site is small this isn't the end of the world; this becomes more and more of a concern as the site grows.
- Receiving a notification if your comment was marked as such helps, but does not fully alleviate my concerns.
- (If you get a notification on an account you no longer check, that doesn't really help.)
- (Ditto, if you get a notification that you disagree with... then what?)
- For sorting: either people ended up upvoting said comment a lot - in which case knowing that e.g. 'this post was heavily edited to address X and many of the comments are discussing pre-edit' is useful - or they didn't, in which case there's no real need to downsort the comment.
There is already a user setting "Do not collapse comments to single line", it makes sense that this would also apply to editorial comments.
I would be fine if this was combined.
- ^
The post starts off by talking about typos, but then says "this comment was useful at the moment it was made, but it has outlived is usefulness" - which covers far more than just typos.
- ^
The current moderation is bad enough for this; it's a whole lot easier to detect bad-faith 'removing comment entirely' than it is to detect bad-faith 'I did an edit that (didn't actually) address(ed) this point, so this comment no longer applies'.
↑ comment by Viliam · 2022-04-16T11:51:29.238Z · LW(p) · GW(p)
I would have far less objections if the comment author was the only person that could mark a comment as such. Unfortunately, I recognize that in the good-faith case this somewhat reduces its utility.
Yeah. I was considering something like "the 'outdated' flag can be set by comment author, article authors, or the moderators... but if someone else has set it, the comment author will be notified and can remove the flag, in which case no one other than the comment author will be able to set it again" -- but this seems needlessly complicated.
From my perspective, in the case of abuse, the comment is not removed, only collapsed and moved to the bottom, so there remains a clear evidence of abuse that can be called out; which in my opinion is a sufficient protection against abuse. (I guess there should be a small print mentioning who specifically has set the "outdated" flag for given comment.)
Maybe the wording should be specific enough, like "This comment reported a typo which has been fixed", to make it more obvious what constitutes an abuse of the button.
Replies from: TLW↑ comment by TLW · 2022-04-16T12:36:03.418Z · LW(p) · GW(p)
Yeah. I was considering something like "the 'outdated' flag can be set by comment author, article authors, or the moderators... but if someone else has set it, the comment author will be notified and can remove the flag, in which case no one other than the comment author will be able to set it again" -- but this seems needlessly complicated.
(If you get a notification on an account you no longer check, that doesn't really help.)
so there remains a clear evidence of abuse that can be called out; which in my opinion is a sufficient protection against abuse.
Hm. I think you have a somewhat more optimistic view of this than I do.
I do not believe that "a (failed) attempt to fix an issue" can be that easily distinguished from abuse.
It's not so much "this comment reported a typo". It's things like "step 6 of your argument doesn't follow from step 5" hidden and the post updated with something that doesn't actually resolve the problem.