MathOverflow

12:00 PM

Typically, closure is accompanied with one of a few canned messages, e.g., the one mentioning Math.SE as a site for general questions in mathematics, and no other accompanying comments. Is that an example of intolerance or use of an iron fist? I'm trying to get an exact idea of what you mean. — Todd Trimble ♦ 12 mins ago

@ToddTrimble I suppose fedja will respond, but the way I understood the post is that sometimes MO folks seem to be over-eager in closing questions.

One example, which I know of simply because I interacted with fedja a bit, can be seen here: chat.stackexchange.com/transcript/19138/2017/12/1

And quite recently you and Fedor Petrov have reopened this question, after it was initially close: How to evaluate this integral?

I will admit that my comment, with a link to math.SE post, might have contributed to that closure. (Only after posting the comment I've noticed that the question was about the integral and not about the limit. However, knowing the limit probably still counts as useful enough to be posted in comments.)

I suppose that similar discussions have been around long before MO joined SE network. For example, I see Elitism for elitism's sake on tea.

And I guess that, to some some extent, it is intended that MO should have elitist feel.

BTW How to evaluate this multiple integral? was relatively quickly close on math.SE and then deleted by the OP.

Todd Trimble

12:29 PM

@MartinSleziak Thank you, Martin; that was very helpful indeed. I often have the same feeling as fedja, as I commented under a recent meta post of Asaf's.

Martin Sleziak

I see.

Hear, hear. This problem has come up numerous times, on the other side as well where non-expert reviewers also vote to close when they shouldn't, using some superficial criteria. I think it's the latter which bothers me much more, since it discourages new people from wanting to participate. — Todd Trimble ♦ yesterday

I see that fedja also responded to your comment on his question.

We will see whether something will come out from that discussion.

Martin Sleziak

1:34 PM

@MartinSleziak I agree with you that incorrect migration is just as bad as an unwarranted closure. Indeed, what was the point of migrating this one mathoverflow.net/questions/282730/… say? Note that the migration was done on the same very "not research level" basis I abhor so much, while the real issue with the question is what I pointed out, say, here mathoverflow.net/questions/288411/… Do you agree with this assessment? — fedja 5 mins ago

As I am below 10k+, I do not see the question proof for mini batch mini-batch gradient estimator linked in the previous post.

Still, after a quick google search, it seems that it was copied by some scraper sites before deletion.

Can somebody with sufficient reputation confirm whether this is indeed the complete content of the post?

> Suppose a mini batch size m selected uniformly from size n. And a loss function L(x, y, θ), θ is the weight to be optimized. How do I prove that the excepted avg gradient for this loss function is the same as the true avg gradient for n?

Should I be able to see at least the migration-stub on math.SE? (I have 10k+ there.)

Was it actually migrated to math.SE or to stats.SE?

I can see at least part of that post in Google Cache here.

So it seems that it looks the way I described above.

@fedja Since only 10k+ users can view deleted posts, I don't really see what your "batch" example actually said. If the contest is actually the same thing I was able to get from Google Cache, then I think it was reasonable to expect that the post is going to be closed on the target site. To me it seems that post with such content should be closed (but not migrated). — Martin Sleziak 9 secs ago

@MartinSleziak And this mathoverflow.net/questions/288478/the-sobolev-space-w1-4 looks like a sure migration to me (if you disagree, let me know what is wrong with it for MSE). Let's see how it will end up (I'm not voting) — fedja 1 min ago

1

Q: the sobolev space $W^{1,4}$

Let the function f be: $$f(x)={\left(1-\frac{2|x|}{X}\right)}^{\alpha}, $$ X is just a real parameter. How can I find the set of $\alpha$ such that $f \in W^{1,4}([-\frac{X}{2},\frac{X}{2}])$ ? I know the definition but I'm new with this kind of question so any help is appreciated.

sobolev-spaces

It seems that the above post is still in the close votes review.

If the question was posted on math.SE, I would certainly leave judging whether or not it is suitable to people who know about Sobolev spaces.

Martin Sleziak

3:00 PM

@MartinSleziak Yep, it is. And any person who would care to think of whether the question is clear or not instead of thinking whether it is up to his standard of research level or not would vote Close(Unclear what you are asking) instead of Migrate(Not research level). So IMHO it is still the same attitude I lament about and not a system defect or MO/MSE standard differences that leads to high percentage of migration rejections you are unhappy with. — fedja 2 mins ago

@fedja It seems to me that we have digressed from your original problem to question what should and what should not be migrated. If we want to continue this discussion, maybe it would be better to do so in chat, so that we do not leave here too many comments that are unrelated to the issue at hand. — Martin Sleziak 9 secs ago

I will add this very recent question as an example of a post which should be closed but not migrated: Discuss the continuty of a real valued function.

It would be perfectly fine to suggest to the OP to ask on MSE. But the OP should be advised to improve the question before posting there. (And, in this case, probably also to look for duplicates instead of asking the question.)

BTW the "batch" example was migrated to stat.SE and not to math.SE, right?

fedja

After you handled it the way you did, I agree that pure closure is appropriate. As to the batch example, you are, probably, right. I rather cared about whether to send it out at all than about where exactly to send it. As to the digression from the original theme, yes and no (see my last comment in the main thread)

Martin Sleziak

I am below 10k points, but you should see in the revision history to which site the post was migrated.

The point I am trying to make is that the question about Thomae's function would be quickly closed as off-topic on math.SE, so it does not makes much sense to migrate.

fedja

Yeah. stat.SE.

Martin Sleziak

Although it probably does not make that much difference whether it stays here as a migration stub or as a post which was closed as off-topic.

fedja

For the OP, no. For the relations between MO and stat.SE it does. As to the Thomae question, I'm not sure. I've seen much more trivial stuff on MSE that was happily upvoted and discussed for a long time.

Martin Sleziak

3:13 PM

About your Sobolev space example - I am definitely not qualified enough to judge whether it should be closed on MO. And whether it would be closed on math.SE - it is pure speculation.

@fedja The Thomae question would not be closed because it is too trivial. It would be closed with the close reason "missing context".

Generally, posts which contain nothing beyond a statement of the problem are closed on math.SE.

I agree with the following:

> IMO ideally anyone casting a "no effort" -close vote should be able to solve the problem themself. I am aware that the policy I suggested may place too high a burden to the first close-voter. That's where that ideally came from.

It is a quote from this answer by Jyrki Lahtonen.

fedja

It is a relatively standard exercise in the introductory course in Sobolev spaces. Just like figuring out the set of values of $a$ for which $(x^3-a)/(x-a)$ is bounded near the origin in a calculus course.

Sorry, not "near the origin", but, say, on [0,1]

Martin Sleziak

On MO, the Sobolev post got so far one vote to leave open and one vote to close in the review.

I am not sure to which extent you are familiar with PSQ/missing context disputes on Mathematics Meta.

By far the most frequent close reason on math.SE is currently missing context.

Here is one example: Complete and compactness of metric space.

You can see the banner under the question saying:

> This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level.

fedja

As I said, let's see :-) Not at all (for missing context) but I'm quite opposed to this formulation too in general. Still, it can be perceived as an "improvement request", so it is less unreasonable

Martin Sleziak

The text "improve the question" links to: How to ask a good question?

All I am saying that anybody who is familiar enough with current way question like that are handled on math.SE can immediately see that Thomae question would be closed either as missing context (or as a duplicate) unless the OP adds something more to the question.

IIRC closure and duplicate does not mean that migration is rejected. But closure as off-topic/missing context would lead to a rejected migration.

At least I will ask this. You're post on meta suggests that handling such question on MO is too harsh and the comments are unfriendly/arrogant.

If Discuss the continuty of a real valued function and the comments are the way they are now, would that be ok, or is it still unsatisfactory?

fedja

As to Complete and compactness, what "context" do you expect? It is crystal clear that a student has a struggle with a textbook exercise and even the notation is somewhat confusing to him. Do you want to know which edition of Apostol or something else it is and what is the page number, or what?

Martin Sleziak

3:25 PM

I'd assume that you in your last message means: you = math.SE community and not you= Martin Sleziak.

fedja

Correct :-)

Martin Sleziak

I seldom vote to close, since to me it seems that there is not clear consensus about what should be closed and what should not.

fedja

You handled the Discuss the continuity one as neatly as one can possibly wish for. So closing now is totally fine.

Martin Sleziak

There are some guidelines in How to ask a good question/Provide context.

If I look at the question I linked as an example: Complete and compactness of metric space

At the very least the OP should include what metric is used on C(X,Y).

fedja

Yeah, so which one is applicable here in your opinion?

Martin Sleziak

3:29 PM

Re: Do you want to know which edition of Apostol or something else it is and what is the page number, or what? Adding a source is certainly appropriate way to add context.

And that would also count as: "Indicate your own background".

fedja

Correct. And that was asked and it is clear from the answer that OP is somewhat confused about this already. As to the source, that may help a bitin this particular case, but if the question is clear and self-contained, what difference does it make?

Martin Sleziak

Well, to be blunt, one important difference is that the question has better chance not to get closed.

fedja

Sorry, my typing speed is 3 times slower than yours :-) He did indicate his background very clearly. The words "I have no idea and am new to this" have meaning, don't they?

Martin Sleziak

Another important thing is that Math.SE should not become "homework mill" where everybody can copy-paste a homework assignment and get it solved. (This is exactly the reason why many people here at MO snub Math.SE.)

So to expect at least some very minimal standard from the question is not that unreasonable.

fedja

There is a difference between a "homework mill" and a "help site". MO is also just a help site (the way I see it).

Martin Sleziak

3:35 PM

I will probably just quote this:

I think it is somewhat misleading to think the difference between MO and MSE mostly a difference of cultures. Much of the structure required at MSE is a necessity that comes from it being the second largest SE site and by being frequented by drive by users trying to dump their homework[...] — Michael Greinecker Oct 24 at 17:18

[...]The MSE meta lingo for problems lacking any context is PSQ for "problem statement questions", and this notion came mostly out of discussions on how to deal with homework questions. The need for context also arises because one may want to help a specific asker. You don't often read "Sorry, we didn't learn that concept yet." on MO. — Michael Greinecker Oct 24 at 17:18

I am in slightly weird position here.

I am not big fan of closing questions. (But I also understand the point of people who are fighting against lazy question being dumped on that site.)

So I am kind of arguing the case for a side I don't completely agree with.

fedja

You don't often read "Sorry, we didn't learn that concept yet." on MO See my (8th from the top) comment here mathoverflow.net/questions/288191/… :lol:

Martin Sleziak

This one?

@user2478159 I think that if you spend some time figuring out exactly what you really want to ask and post the question in a new thread, I'll think of it :-). Patchwork is never good in my experience though we still do have some leeway in the current argument. — fedja 14 hours ago

fedja

But I agree that there is a fine line between "dumping homeworks" and "asking for help when stuck".
Nope, above it (3/2 volumes of Hormander)

Martin Sleziak

@user2478159 Just write $F(3\cos x)=\sum_{k\ge 0}a_k\cos kx$ (Fourier) and put $P_k(3\cos x)=a_k\cos kx$. Note that $a_k$ decay faster than any power and $x\to 3\cos x$ is a diffeomorphism of some arc to $[-2,2]$, to get the bounds for $P_k^{(m)}$. The estimate $|P_k(z)|\le C_{0,0}(1+|z|)^k$ is just the maximum principle for $G(z)=P_k(z+\frac 1z)z^{-k}$ outside the unit disk. As to analytic front sets, I managed to read only $3/2$ of $4$ volumes of Hormander when I was a student, so I don't understand the words. However, if you translate it into elementary language, I can think of it. — fedja yesterday

fedja

Yes

Martin Sleziak

3:42 PM

If that was meant as a rebuttal to Michael Greinecker's comment, this is slightly different. "We didn't learn that yet" does not come from the OP in this case.

Re: what "context" do you expect?

This is something I quite often struggle with.

It sometimes happens that there is a question which has interesting answer and which might be useful for other users of the site.

fedja

It is an exact equivalent of "Sorry, we haven't learned this concept yet". And I can bring any OP who asks an analysis question on MO to this state (well, probably Terry should be excluded) if I choose to answer in the language I know. It is OK to request to translate into common tongue on either side

Martin Sleziak

Still, some users insist on closing it because of missing context. And sometimes it is quite difficult to think of something to be added in order to satisfy the requirement to include context.

fedja

The requirements are guidelines, not a dogma.

And the "missing context" one is more like an excuse to close than a real objection. What exactly does one need to know the context for?

Martin Sleziak

I am not sure whether it makes much sense to reiterate the discussions which were held again and again before.

For example, if you look at arjafi's answer to a related post, you will notice some familiar faces.

There aren't many things as frustrating as writing a good answer to a question without context, and then the OP comments "I don't understand any of these concepts". Of course, the main goal is to help others, but it is much easier when you know what the OP is after. You can even have several posts with the same question, but with very different answers, if they are calibrated to different levels and techniques. So when you ask someone to show their work, they have to reveal what they know about the problem, which can give a good estimation as to what sort of answer is sought after. — Asaf Karagila May 6 '16 at 19:05

Why don't we make an experiment which is at least a bit constructive instead?

My suggestion would be - let's try to improve the C(X,Y) post at least a bit to see whether it gets reopened.

Definitely, including definition of C(X,Y) and which metric is taken on that space would be an improvement.

The title can be made a bit clearer.

fedja

Go ahead. It is on your side now. I tried once on MO, remember? :-)

Martin Sleziak

3:54 PM

You mean "Nonlinear Differential Equation System Example", right?

fedja

Yeah

OK, I have to go now. We can continue later, if you wish :-)

Martin Sleziak

4:07 PM

I have edited the post and I have explicitly mentioned in the edit summary why it was edited by somebody else, not the OP.

My guess it that the post is not going to be reopened.

But at least I tried to improve the post in the way I was able to.

Re: There is a difference between a "homework mill" and a "help site".

I do not view math.SE primarily as the help site. By which I mean that it is probably more important whether the Q&A pair is useful for other users who display the same question later than whether it is useful to the OP.

To me it seems that the OP would, in this specific case, probably benefit much more from the conversation about the question in chat. That is much more interactive and the OP would be able to ask about specific steps which they do not understand.

However, this whole discussion started from the fact that question which are simply PSQ are almost always closed on math.SE.

This is simply the fact. Perhaps we don't like that this happens, but this does not change that the site works this way.

Martin Sleziak

4:22 PM

Re: The words "I have no idea and am new to this" have meaning, don't they?

Something along those lines was discussed here: Suggested Guideline for “I Don't Know Where to Begin” Questions

Quote from one of the answers:

If I had one of these three people in my office, I would ask them to write down the definitions, and we could proceed from there to have a discussion about it. On this web site, the users most in need of this kind of back-and-forth discussion are those least likely to participate in one, so I don't think we can usually help them. In the past I have often left comments....

... 1 2 3 4 ...

suggesting that the writer should start by writing down the definition of compactness; these comments are almost never answered. (Item 3 is a happy counterexample.) My current practice is usually to downvote the question (as a warning to other readers) and move on quickly.

Anyway, I should probably be doing something else, too. Have a nice day and good luck fighting with closures!

I am a bit sorry that this exchange wasn't more optimistic.

Martin Sleziak

5:00 PM

So the experiment did not work out as we wished - the reviewers considered the improvements insufficient for the post to be reopened.

fedja

10:09 PM

To me it seems that the OP would, in this specific case, probably benefit much more from the conversation about the question in chat. That is much more interactive and the OP would be able to ask about specific steps which they do not understand. This is always an option and I use it now and then (see here mathoverflow.net/questions/288136/… for instance).

Takes my time, of course, but it is my problem. So my response to There aren't many things as frustrating as writing a good answer to a question without context, and then the OP comments "I don't understand any of these concepts". would be "perhaps, but why do you care? The person who answered should decide what to do and he can end up with making a snarky remark and voting to close, but what entitles him to such behavior is his giving an answer in the first place. And what entitles you?

OK, going home now. It was quite a long day. Thanks for the attempt to edit that MSE post, by the way ;-)

Transcript for