8
$\begingroup$

As explained here, I asked a question on Stack Overflow. A user with high reputation suggested me to move the question to Cross-Validated, and since in meantime the question had received 3 votes to close it and move it to Cross-Validated, I decided to delete the question and move it to CV here. It was put on hold immediately, in less time than it would take to me to read a question that long, without a specific comment indicating what precisely I could modify in order to improve the question. I see that it is not the first time this happens. So what should I do now? I frankly cannot believe the question doesn't belong anywhere on Stack Exchange. Can I get some useful comment helping me to improve it? If I have to move it again to Stack Overflow, what can I do to avoid getting new requests to move it here?

$\endgroup$
  • 3
    $\begingroup$ I think the q. sends mixed messages: asking for an explanation of what's going on when you try to use RBF with 5 centres is on-topic; but "I need an answer which uses R" - "gimme teh codez"! - is a clear off-topic flag. A slight edit to favour the former reading will probably improve the chances of its being re-opened. $\endgroup$ – Scortchi Sep 9 '16 at 14:16
  • 2
    $\begingroup$ while I like your angle, I don't think my question is a polite way of writing "gimme teh codez" :) I would just need an hint on how to fix my code, or whether to use another package in case RSNNS is buggy. I cannot actually ask "which package should I use for RBF interpolation in R" because that's against the policy. So I tried to 1. find a package myself, looking on the Internet, 2. learn the documentation, 3. make some tests myself. I think this can be considered "due diligence"...If you still don't agree, I will make the update you suggest. $\endgroup$ – DeltaIV Sep 9 '16 at 14:33
  • $\begingroup$ ps note that nowhere in RSNNS documentation there's a hint to the fact that function rbf cannot be used for interpolation. As a comparison, lm will happily fit a model with N parameters to a dataset of size N+1, and simply warn you that you have zero residual degrees of freedom. $\endgroup$ – DeltaIV Sep 9 '16 at 14:37
  • 4
    $\begingroup$ Sorry - I was being rather flippant, but wanted to describe how the question might be perceived. It's often the case that once the statistical question's answered, how to implement a solution is the easy part. $\endgroup$ – Scortchi Sep 9 '16 at 14:49
  • 3
    $\begingroup$ I would agree that you have done your due diligence. Reading the question again, it still doesn't sound like you are asking about how interpolation or RBFs work, or some other statistical / ML question. It sounds to me like you are having trouble with the code. It really seems like an on-topic question for Stack Overflow. I would not have voted to close there. $\endgroup$ – gung Sep 9 '16 at 15:23
  • 2
    $\begingroup$ @gung: I think there's a statistical (or computational) question in there - why doesn't adding enough centres to an RBF fit result in interpolation, as adding enough polynomial terms does in regression? Depends what DeltaIV really wants answering. $\endgroup$ – Scortchi Sep 9 '16 at 15:32
  • $\begingroup$ @Scortchi, that would be on topic here, IMO. I didn't recognize that that was the question. If that were clarified, I'd be happy to vote to reopen. $\endgroup$ – gung Sep 9 '16 at 15:38
  • 3
    $\begingroup$ I will make the modification suggested here and hope that the questions will be answered, but frankly the moderation by @PeterFlom leaves a lot to be desired. If he really believed the question was a programming question, he should have had that coherently migrated to Stack Overflow. And that would have been a lot of fun, given the precedent... $\endgroup$ – DeltaIV Sep 9 '16 at 15:38
  • 2
    $\begingroup$ +1 here, there, and voted to reopen, but to be precise: your Q was closed 42 seconds after it was posted, not 1 second after. $\endgroup$ – amoeba Sep 9 '16 at 20:08
  • 3
    $\begingroup$ The question hit the reopen queue very quickly after closure (I assume you edited it) -- and in turn a number of our high rep users voted to keep it closed. As a result, even if the decision were later widely regarded to be a mistake (&it is not at all clear that it is), it's not reasonable to lay that at Peter's feet; a number of people able to reverse the choice agreed with him, and would presumably have closed it themselves if he had not. As a result, while it's reasonable to discuss the merits of the decision, I don't think it's okay to make it about one person, since clearly it was not. $\endgroup$ – Glen_b Sep 9 '16 at 23:08
  • 5
    $\begingroup$ In addition, the best thing to do from your post on SO would be wait for the post to be migrated, not to post a second copy. One advantage of that is that if SO voters are mistaken about it being on topic here, there's at least some signal of it back to SO. $\endgroup$ – Glen_b Sep 9 '16 at 23:14
  • 1
    $\begingroup$ See the third-last paragraph here. The advantage of migration is that you don't end up with a closed question at the source site waiting for improvements that won't happen (since it's deleted after migration). The "time and effort" to migrate is a few clicks, and it doesn't take a moderator to do it if there are high reputation users. Small sites (where only the mods can do it) generally don't find it a burden. If you delete the original yourself it's not a big deal either way, but most people don't ... ctd $\endgroup$ – Glen_b Sep 11 '16 at 9:59
  • 1
    $\begingroup$ ctd ... and then when the high rep users migrate the original, we end up with both copies. Getting two copies is more of a hassle because then you have to notice the duplicate before closing one. And if the first one gets answers on the original site before it migrates, while the second posted copy on the new site gets answers, you then have to worry about merging. One the whole, migration is a lot less hassle than dealing with that. $\endgroup$ – Glen_b Sep 11 '16 at 10:16
  • 1
    $\begingroup$ @Glen_b, ok, understood. I couldn't have known all that. If I'm told it takes time and effort, I try to be of help. Hey, about the third-to-last paragraph you link, if that's an invitation to let communities migrate questions, then ok, I got it. But since that paragraph also mentions cross-posting, I want to make it clear that I rigorously did not cross-post. I know it's frowned upon on Stack Exchange, so I deleted the question on the source site before posting on Cross-Validation. $\endgroup$ – DeltaIV Sep 11 '16 at 10:29
  • 2
    $\begingroup$ Yes, you didn't cross post (and thank you for that); hence my explicit "If you delete the original yourself it's not a big deal [...] but most people don't" (I wouldn't have included that if you had not done so). And I agree you weren't necessarily to know all that stuff, which is why I tried to explain it -- that was the original purpose of my comment about migration... to give more information. It's not that you did the wrong thing there. $\endgroup$ – Glen_b Sep 11 '16 at 10:56
6
$\begingroup$

If you consider the universe of questions which are daily closed or reopened it will not be the first nor the last time to happen. It just happens.

If you spend some time in this community you'll see how our experienced users are engaged to provide feedback and help reversing situations like that.

Remembering that letting a comment after putting a question on hold is not mandatory (despite a comment can always help), because when a question is put on hold there is already a notice message explaining why.

Let's go to your questions:

So, what should I do now?

First thing, edit the question to address the notice message (and comments, if any). In your case, edit the question and explain why it is about statistics and not solely about programming. The question will automatically go to the Reopen queue. There is a chance reviewers might decide to reopen it.

If it does not (and OP still believes he/she has a case), asking for help on Meta (or chat) is always an option (like you did).

Can I get some useful comment helping me to improve it?

Yes you can, but you need to take advantage of the hints you receive. In my humble opinion, you could have explored more Scortchi's advice. So, I am suggesting another edit to your question. See if you like it.

If I have to move it again to Stack Overflow, what can I do to avoid getting new requests to move it here?

I'd say edit it to emphasize it is a pure programming question, rather than involving statistics knowledge to get it solved.

$\endgroup$
  • $\begingroup$ To be clear, when you say that "it just happens", you mean that it just happens that a question about to be migrated to another site, is closed on the new site, right? You are not talking simply about questions being closed. Because of course here the problem wasn't simply that the question was closed. I had questions closed on single sites in the past, and I didn't have a problem with that. $\endgroup$ – DeltaIV Sep 11 '16 at 10:19
  • 1
    $\begingroup$ I didn't notice you already edited my question. Yes, I think it's ok. $\endgroup$ – DeltaIV Sep 11 '16 at 10:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .