Handling "misvoted" content

When I first got privileges to go through review queues, I did. A little bit. And then I found they were really tedious. Boring. Waste of time. Especially if I am really into a particular site where I’m reading nearly every post anyway. On a really big site (SO) they may be necessary because of the volume of posts - but then you are going to have things even worse because there is just too much stuff.

So if we have review queues they need to be different from SE. Not sure how, but the SE setup just doesn’t work for me.

“Menial work” is OK, when I can pick and choose. I will often make little edits (grammar, spelling, etc.) and sometimes a lot of edits on one answer (there is one expert on Retrocomputing who could really use a proofreader - but his answers are technically excellent so worth the time for me to clean them up to make them better for everyone).

3 Likes

Totally agree. I had the same experience.

Perhaps different mechanisms are needed to really large sites. It always pissed me off that SE seemed to treat everything like it was SO.

One thing I suggested on SE (which was ignored) is to put a “reviewed” button on each post. Like you, I read most questions anyway. When I’d occasionally look in a review queue, I’d see lots of posts I’d already looked at separately. That was annoying and resulted in a “screw this” attitude from me.

Give me a way to mark that I’ve looked at something and taken all actions I’m going to take. That way they won’t show up as needing review by me, and should result in a lot less in the review queues. If there are only 4 items in a review queue, I might go clear them out. When there are 40 it feels too much like bailing the Titanic with a teaspoon, and I’m not going to bother.

5 Likes

No! Editing is absolutely essential. Any system that doesn’t allow free editing is basically pointless as far as I’m concerned. The single most important feature of the SE system, for me, is the collaborative editing. If we don’t have that, we may as well not build anything at all and use existing forums.

5 Likes

Well, that may be true of you, but it certainly isn’t true globally. Many, many experts absolutely do review. I am active on several SE sites and see this everywhere. So just because you don’t enjoy the janitorial work (and there’s no reason why you should) doesn’t mean that “all experts” don’t enjoy it. Not by a long shot.

But that’s the beauty of a system like what we want to build here: we each do the bits we enjoy doing. I have done thousands of edits over the years, for example, because I enjoy it.

8 Likes

I wasn’t objecting to editing capability. I agree that is good, and works reasonably well on SE.

What I was disagreeing with was your suggestion to edit a post to say it is wrong and where to look for the right answer.

To me it is completely wrong to edit someone else’s post to change the original intention.

4 Likes

I hear you, but I personally strongly disagree. If a post is wrong or dangerous, editing in a warning seems like the best choice. I don’t really care about the original intent of the poster, I think we should always aim to create a library of information and shouldn’t care much about who wrote what or how.

So yes, avoid editing just to change the style of the post so that it is how you would have written it, but if the post is wrong or dangerous, editing it to make that clear seems fine.

4 Likes

The problem is that calibration of “wrong” and “dangerous” varies. In every tech (or other) religious war, some assert “wrong” where it doesn’t objectively apply. Who’s going to be the gatekeeper?

I agree that “well, write a competing answer” isn’t always the solution either; an expert should be able to express that something is wrong without having to also explain what the right thing to do instead is. (That might deter the former, at greater public cost.) So I want a way for that expert to say “this is wrong”, but not by putting “this is wrong” in the mouth of the person whose name is on that answer, a person who might believe just as strongly that it’s right and the complainer is arguing over a religious matter rather than something more objective.

Really, instead of either “right” or “wrong”, I want to see people explain the consequences of an answer – if you do that then you’ll damage this in this way, etc. The author might then say that eh, that cost is minor compared to such-and-such benefit – and then you have a list of considerations rather than just a value judgement, and people can evaluate that against their own priorities.

6 Likes

If an expert upvotes it, doesn’t that give this information already? We can do something like this behind the scenes if we want (which may just involve sorting it differently, perhaps showing a “quality score” percentage in addition to the vote count) without having to worry users with having essentially another way to upvote.

The only reaction I would propose would be a “user went above and beyond” reaction, which is also essentially another way to upvote, but it’s not “this is right”, like the above, but rather “this is amazing”, which would be a nice way to give an extra reward to posts (even if it doesn’t affect score or any type of reputation) and might only need to have a very basic contribution requirement. Although limiting how many you can use in some way would make sense.

I think there should be a way to add a notice to answers (for example code block solutions) which have become (or turned out to be) dangerous, in form of a special prominent warning notice linking to an explanation by the user who added it. I think it could be a checkbox in the comment editor. To add such a notice you’d have to also provide a url source or a quote with the source’s name or identification, so that it could be later checked by reviewers.

Such a comment’s formatting could resemble a list of citation sources like on Wikipedia - comment’s text block and below it a numbered list of individual sources to support the claim that the answer is dangerous.

Because sometimes sites get big and not 100% of its content is under constant surveillance and review, I think it would make sense to give the poster the benefit of the doubt and show such a warning notice by default, albeit with a slightly modified “Not confirmed by other users yet” connotation.

Example of such a notice at the top of the allegedly dangerous answer:

“Warning! Potentially dangerous solution. Please read the orange comments below before using the described solution.”

Reviewers, moderators and “gold badge owners” can then vote (vote breakdown should be always visible and public, I think, if we’re going to want to trust the votes) on each individual supporting claim to verify or reject claim sources. If all sources are rejected, the notice is removed, but the comment can still stay. If at least one source is approved, the notice stays but with indication of this fact. Prominence should only be reduced when all the sources are invalidated, in my opinion, only the description of how definitely dangerous the answer is should change.

Approval and rejection of these claims can be done as a sub-comment to the notice-adding comment. For example:

  • (verify dangerous claim) This is indeed dangerous, as per the updated spec and this specific quote here: [quote from spec]
  • (verify even if with a bit of doubt) This is only dangerous if you’re running the service as admin, which you should never do. Nevertheless, a lot of people run it as admin, so this should stay up as dangerous.
  • (invalidate dangerous claim) This is not dangerous because this situation can literally never happen on a modern CPU, and the app does not run on the old CPU at all. If it could, that claim would be valid.
2 Likes

The idea is to have a mechanism to partially counter the effect of pure votes, so that an expert can somehow convey :warning: “this is popular (lots of upvotes), but wrong (trust me, I’m an expert)” or :+1: “this has been ignored (very few upvotes) but it’s actually great (so says I)”.

This feels a little overcomplicated. I’d prefer to keep it simple: a reaction plus a comment, and for egregious cases, edit in at least the information that the answer is contested with a link to a counterpoint.

4 Likes

Sure, we could start with a simple version and complicate it if it’s warranted later.