Should we delete answers that don't meet the requirements of the hard-science tag?

Question

Around the beginning of August, Stack Exchange created a custom post notice for Worldbuilding. Basically, moderators manually add the notice to all questions tagged hard-science, and then may add it to specific answers that don't meet the requirements of the tag. There are only four of us, so it sometimes takes a day or so for the post notice to be added, but in general, it gets done.

I've monitored posts on the tag over the past month, and we've had to add the post notice to a bunch of answers - probably 50% or more, though I don't have the exact statistics. Adding the notice seems to have some effects, but not many. hard-science answer quality is, for the most part, staying the same. I'm not saying that there hasn't been any improvement - I've seen some really awesome answers over the past month - but it hasn't been as much as I think many of us hoped for.

That said, I'd like to propose something that may be unpopular, but that I'd like to put on the table for discussion: Deleting answers. To my knowledge, this was proposed early last summer, though we never took action in that direction. At the time, it was too extreme a response. Now, though, I feel like it might be a better choice. We've tried many things, and none of them seem to be working. Here's my proposed solution:

1. If an answer to a hard-science question doesn't fit the requirements of the tag, it can have the post notice added to it.
2. If the answerer doesn't improve the answer within $X$ days (where $X\gg1$), the answer may be deleted.
3. If the author improves the answer, he or she can flag the post and ask for it to be undeleted, and to have the post notice removed. Hopefully, this will be the case.

This proposal has some problems and questions we need to answer, as I think people have figured out by now:

• Who is/are the arbiter(s) of whether or not an answer meets the requirements of the hard-science tag? Furthermore, the issue of whether or not an answer satisfies this has recently been brought up in Does my answer fit the hard-science tag?, though there's been neither answers nor a consensus.
• How much time should the answerer be given (i.e. what is a reasonable value of $X$)?
• How can we easily keep track of which answers have been "given notice", as it were, and when they may be eligible for deletion?
• Is there a way to bring this into the infrastructure of Stack Exchange? What if an answer has a net score of +5, and might not be sent to the Low Quality Posts queue? Might a moderator have to deal with that? How can we get community consensus (see also Bullet #1)?
• What if there are disputes as to the deletion of an answer? If the answerer feels the deletion was unfair, how can this be solved, given that many people won't be able to see the deleted answer?
• Would this solution cause more problems than it solves? Is it worth potentially starting disputes for the sake of scientific accuracy?
• Am I just being overly persnickety?

These are all very potential points against the idea of answer deletion (though none have simple yes-or-no answers), so let's be clear about something: I'm presenting this proposal from a neutral standpoint. On the one hand, there are quite a few problems, obstacles, and roadblocks when it comes to implementing this. On the other hand, this would be a way of dealing with low-quality answers; it might not be the optimal way, but I'll be honest: I'm all out of solutions.

Answer 1

So my view on the subject.

Who is/are the arbiter(s) of whether or not an answer meets the requirements of the hard-science tag?

The community.

Instead of taking the step of deletion, I think we should first have a stamped-out comment that people can post, something like:

"Hard-Science tags require answers that _______. This answer does not appear to abide by that, and thus the asker's request, and thus is a low quality answer. Consider revising or removing this answer."

(Somebody else can write it better than that, but the idea still stands)

The community will, theoretically, either stop upvoting the answer/downvoting it (as 'dangerously incorrect' as it is not the hard-science requested), on this question, and slowly, hopefully, a culture will form... or they won't, and they'll continue to upvote, ignoring the tag.

If this happens still, we reach a logical decision point:

1. The community might agree with the idea if they knew, but they haven't gotten it through their heads. They need to be--ahem--further enlightened.

...In this case I'd escalate to the suggested deletion option, although this seems significantly more major and heavy-handed. We should look at this option at that point down the road. Jumping straight to deleting answers seems a few steps too fast.

2. The community by large disagrees with the idea, and they community is wrong! Dead wrong!

Gosh darn it, the World Building community doesn't know what's right from wrong and keeps getting out of line! These whipper-snappers need to respect the views of Meta and...

...No but seriously. If we think the World Building community (including those who don't view meta) by-and-large get the Hard-Science tag and think it's useless, ignore it, and support others ignoring it perhaps it is we who are wrong. My opinion is that it's useful, but if we reach that point--where I stand against the 90% in my down voting soft-science answers, waving my banner proudly as the ship goes down--I have humility enough to say that maybe it's not a useful tag.

If we reach that point, let's not say "it's the 90% who don't know what's good for them."

Answer 2

I'm a moderator on The Workplace, where we have a "back it up" rule. Answers that don't comply, that come to moderators' attention, get post notices, and if a couple weeks go by without it being fixed and that comes to moderators' attention, we delete the answer. (We have a user who pretty regularly goes through posts with post notices and flags them.)

That works well because any moderator can usually tell whether the answer meets the rules. If it's unclear or there's a dispute, somebody brings it up on meta.

Hard science feels different, at least to me. My ability to assess the hard-science-ness of answers could (depending on the subject) be limited to "those links look like they're to scholarly papers" and "hey, math". When either applying a post notice or deleting a post I'd like to have a little more to go on than that. So I usually wait for HDE226868 to handle them, which doesn't scale.

I could spend more effort -- we do have meta posts and a tag wiki with more detail -- but that starts to feel like it's over the threshold of how much investigative work we should expect moderators to do. What I'd like is some sort of clear signal from the community that no, this answer does not meet the hard-science criteria.

There are two cases to consider:

• The post has not been edited since the notice was added. If the community consents, we could rule that after some period of time without objections, the notice is presumed to have been correctly applied and so the post is subject to deletion.

• The post has been edited (beyond stuff like spelling and grammar) since the notice was added. The notice might or might not still apply. We need a way to evaluate these. That evaluation should not come down to a single moderator's judgement.

This is more of a further question than an answer, but it's something I'd like to see actual answers address.

Answer 3

I don't think we should delete an answer if the only problem is that it doesn't meet the hard science requirements.

In such cases they should be allowed to flounder in the usual low vote realms of the additional but not good enough answers that many questions get. Perhaps we should be a bit more liberal with the down votes, but a narrative answer in such cases may also be useful particularly if the asker isn't aware that any "correct" answer would require something like finite element analysis of turbulent flow.

The biggest problem with these answers is that they get more relative attention being on a hard-science question than the 30%-50% of answers on a "viral" question that get few or no votes.

We get very few hard-science questions and worrying about them and the small number of answers they get feels unnecessary. We want to keep the quality high, but I think the voting and flagging system is sufficient without putting additional manual load onto the moderators.

The hard-science flag should certainly remain in place unless the tag is removed.

Answer 4

Nope.

It is inevitable that such a policy would at some point devolve into a screaming internet fight between non-experts promoting their pet models with a a realist furor that mimics religious faith.

SE has a very effective mechanism for controlling the quality of question answers: vote up, vote down. Literally the only time you should be downvoting any question answer is if it does not properly address the question, one example of which is ignoring the requirements of the hard science tag. If the downvoting mechanism isn't providing enough incentive for the answerer to remove their own answer to evade further downvotes, then it should stay. Don't create extra, more central and authoritarian censorship mechanisms. If you feel bad answers are too unkillable, you need to evangelize for diligent down-voting. Many people (self-included) only vote down if someone is actively being a dick.

Answer 5

Let's take look at an example of a hard science question:

How to “set” the lightsail?

Adding pictures, let's call them diagrams, drastically improved the question, but still, is that hard-science demand appropriate for that question?

Probably yes.
Would I like to write and solve all differential equation, accounting relativistic effects? Speeds are not high, but still, it's 5% difference for 0.3c, compared to classical mechanics solutions, totally not scientific not to take this in to account. I do not work with relativistic equation on an everyday basis and the last time I worked with simple differential equation was a year ago. (I had a celebration after solving it.)

But the situation in the question is simple, so the general solution describing all possible solutions in integral form should be simply enough.
And by simplification of functions describing laser power (let's say some constant) will give us some simple solution in a general form.

But based on the question, and some his answers, someone has to invest a huge amount of time to explain the results to JDługosz, or prepare general solutions for his particular case.

General solutions are pretty useless when you can't manipulate with them, also they do not always exists in analytically form (famous n-body problem).
And one way to solve that problem is to narrow down the set of solutions by setting particular initial conditions for searching solutions and by specifying what we are searching for. For example the minimal time, minimal energy etc. Narrowing the question itself helps in searching the solution.

It is an interesting question for me, not exciting, but still useful in promoting sails, nice to see someone else uses lasers to propulsion and tries to figure out how it could be done.

But do I wish go through all the mess mentioned above (even if the hardness of it is a bit exaggerated), when it needs to investigate only 2 cases, a bit of geometry, Newton's third law, simply as that? I would do that for myself, had I such question and would call it good enough for figuring out what to expect in that case.

Another good way would be to make a graphical simulation, although I do not have the proper software for that. (But still, it could be fancy dots moving clip.)

Besides of complains - my point is, hard-science answer should have hard-science question. There is more work in formulating that question than in formulating a good question.

In that particular case the answer is possible, the quality of the question is barely enough to do so but...

Second there have to be people willing to answer the question in such a way, but I do not see why they should choose WB over Physics.SE.

I was nicely surprised by seeing that custom post warning for his question, sad it has to be done manually, but definitely good to see.

Main point, to get his answers we should begin from the starting point, making the question great.

PS
Sorry for not addressing your points, maybe I will edit this post later for that, now I wish to have fun.

PPS
We should have 1,000 points bounties, at least.

Answer 6

It seems to me that it would be more appropriate to look at questions that are tagged hard-science and challenge that request where the question doesn't seem match up to the tag. Many of the questions with the tag I've looked at recently don't seem to understand it, especially from people new to WB, and I've changed it to science-based in a few cases.

Answer 7

What happens if the question is edited to include the hard-science tag after answers have already been posted? Deleting an answer that ignored a requirement from the question may be justifiable (1) (although it does seem somewhat heavy-handed even then; I'm ambivalent). But if someone puts in enough time to make a good answer that is invalidated by a hard-science tag added in a later edit, I really don't think it should be deleted. Downvoted, perhaps, but if it was a good-faith effort that answered the question as it was at the time, it should still have some value. (2) (And we are safe in only considering good answers here, since answers that are both bad and not hard-science will already be voted down enough that we don't need a new mechanism.)

Answer 8

I do not believe we should be deleting or downvoting answers simply because they don't meet meta's definition of what the hard-science tag means.

It appears to me that there are three groups here. (a) OPs with less than 2K reputation. (b) Answerers with less than 2K reputation. And (c) the "Community" which has a fair number of people with more than 2K reputation.

It appears to me that many members of groups (a) and (b) have a definition for hard-science that has more to do with publisher marketing than with group (c)'s opinion.

The problem is that groups (a) and (b) comprise, what, 80% or more of the questions and answers posted to WB:SE? At least it seems that way recently.

Members of the community appear to be trying to impose a definition that is well beyond the expectation of groups (a) and (b) and there is no effective way to educate groups (a) or (b) about that expectation.

Punishing group (b) (and I'm not ashamed to say, me too, even though I just squeeze into group (c)) for trying to help group (a) write a cool story just because neither group (a) nor group (b) meet group (c)'s high standards seems to betray the point of WB:SE — especially when the OP's hard-science tagged question is frequently short of deserving the tag.

This has been a long-winded way of saying:

(1) The average WB:SE user has a publisher's defintion of hard-science and we're trying to hold them to a scientist's definition. That's no bueno.

(2) We frequently punish the answerers when the fault usually lies with the OP's inability to understand what our community thinks hard-science means. Rather than deleting group (b)'s answers, we should be deleting group (a)'s use of the tag.

(3) We deny the OP a lot of really good answers (answers that are often lauded by the OP) simply because we think the tag is more important than user participation.

Thanks for hearing me out. I feel much better and will now return to teasing my cat with a laser.

Answer 9

here's my 2 cents:

• I feel the qualification for a hard science answer can be as lax as having some credible links to statistics and some semblance of a mathematical or scientific process. It being right or wrong is handled by voting.

• What I feel is obviously not hard science is a persons posed unsupported logical conjecture supporting their ultimate answer.

• What I hate is when a question is posed and answered and then revised adding the hard-science tag thus invalidating answers. In this case I do not feel answers should be deleted.

• There is also the issue that sometimes questions do not provide enough information to be reasonably answerable under the hard-science tag.

Thus programmatically I feel like the hard-science tag should only be used under these conditions:

• a question may only invoke the hard-science tag at its creation and not in revision. As this should force the OP to define their question to be conducive with hard science.

• if a question invokes the hard-science tag it should be closeable as not conducive to being answerable under hard-science. Flaggers should also provide an explanation when voting this reason.

• an answer to a hard-science tag should be flag-able for review that informs reviewers that the question needs to present links to facts as well as contain some formulaic construct.

So yes it is in my opinion that posing questions or answers under the hard-science tag should be a mutually rigid and grueling experience that should not be invoked by casual users.

If you want a laborious scientifically supported answer you should put equivalent effort into your question.

I also imagine that there will be an iterative review process between closings and re-openings where readers request clarifications or enhancements of the question.