4
$\begingroup$

My question has just been closed and I can live with it. However, I must ask where it would be better placed. It concerns floating point arithmetic, just as many other questions here. It mentions Java, but any other language giving the same guarantees would do (I don't know other languages in such details).

Anyway, it's not language specific. It's closely related to programming, but it's asking for a proof and thus it's by no means a programming question. An average mathematician could surely do more with it than an average programmer, right?

The fact that I myself answer the question is irrelevant here as I still need someone to look at the proof.

$\endgroup$
  • 6
    $\begingroup$ FWIW, I disagree with closure on this question. I find it a perfectly relevant numerical analysis question. I think maybe some people saw the talk about Java at the top of the question and didn't read further. $\endgroup$ – Emily Nov 17 '12 at 15:03
  • 3
    $\begingroup$ @EdGorcenski I agree, it is a problem in mathematics of computation, so it is on-topic here. There is now a post requesting reopen votes. $\endgroup$ – Bill Dubuque Nov 17 '12 at 18:39
  • $\begingroup$ Now reopened.${}$ $\endgroup$ – Gerry Myerson Nov 17 '12 at 22:19
  • 1
    $\begingroup$ @Ed Gorcenski: Thanks a lot to everybody who voted to reopen. $\endgroup$ – maaartinus Nov 17 '12 at 23:08
4
$\begingroup$

I fee that as the person casting the first vote to close I am obligated to answer. I read the entire question. Twice. Before voting to close. I felt that the question itself was more about the implementation in Java than it was about the numerical computation aspects, which I have to admit I am quite ignorant in. Apparently four other people agreed with me, and five more disagreed and voted to reopen the question.

I suspect that none of the users were harmed, and most likely no one was shot behind their ear whilst walking in the hallways and thinking about the smile hidden behind his mustache.

Now that the question has been reopened and you yourself have posted an answer, I'm sure no one will try to close it again. I should add that a better tagging on your side would have helped hinting me (and those ignorant as me) that this question is about mathematics rather programming.

$\endgroup$
  • $\begingroup$ @Asaf Karagila: My comments are too long so I put them in an answer below. Many thanks for your answer! $\endgroup$ – maaartinus Nov 18 '12 at 18:15
  • $\begingroup$ @maaartinus and Asaf: I have edited to remove a sentence in the original comment and its reply in the answer - both of which might be deemed offensive. Please, let's try to keep discussions as civil as possible. $\endgroup$ – Bill Dubuque Nov 19 '12 at 2:08
  • $\begingroup$ @BillDubuque The sentence still shows up in the answer of maaartinus. $\endgroup$ – Michael Greinecker♦ Nov 19 '12 at 8:36
  • $\begingroup$ @MichaelGreinecker Thanks, I removed it there too. $\endgroup$ – Bill Dubuque Nov 19 '12 at 17:13
5
$\begingroup$

To tackle the "is there a better place?" aspect of this thread, I'd point out the Computational Science SE forum (beta) as a reasonable alternative. Many numerical analysis topics are discussed/answered there. I encourage newcomers and new questions.

Some fairly basic questions about floating point arithmetic turn out to be difficult to treat rigorously, eg. what is the optimal order in which to add some floating point numbers for best accuracy. Of course such difficulty does not mean these questions are not mathematical, only that we often resort to approximate or a posteriori methods for answers out of practicality.

$\endgroup$
4
$\begingroup$

Answer to Asaf Karagila, far too long for a comment.

I read the entire question. Twice.

I must say I'm surprised, but I believe you. Maybe the other four did too and maybe I should've worded the question differently (an advice would be appreciated)?

I felt that the question itself was more about the implementation in Java than it was about the numerical computation aspects...

But even if the question was 100% Java specific... Would it be misplaced here? Note that 99.9% of programmers (no matter in what common language) haven't seen any proof since they've left the school/university, so forget about stackoverflow. Now I've found scicomp.stackexchange.com, maybe it's a better fit and maybe it's not - am I supposed to know???

This is surely not your fault, but each and every border-case question is in danger to be closed, no matter what site you chose.

I suspect that none of the users were harmed...

It looks like it actually helped to attract attention to my question.

I should add that a better tagging on your side...

I'm not familiar with the tags here. Maybe error-propagation or numerical-methods? Feel free to add whatever tags you deem appropriate.

Yes, I'm quite a newbie here (and this isn't gonna change), but what does it say us? My question could have been misplaced here, but then I'd need to be advised or to get the question moved somewhere else. But it was closed and nobody cared to leave a comment; that's not good, especially to a newbie:

  • a real newbie isn't allowed to write comments
  • they don't know how to get a question reopened
  • they might have difficulties to argument in English

I agree that closing a question is a valid way to improve the site quality, but IMHO it gets overused a lot and it happens without comments. Especially in case of "off-topic" some comment would help (both to clarify the situation and to keep the poster calm).

$\endgroup$
  • $\begingroup$ Let me answer some of your questions with indifference, and other with care. I'll be short because I don't want to drag this over too many comments (and surely not into another answer). Do not mistake concise and direct answers as rude, I greatly appreciate your reply. $${}{}$$ I don't know where you should have posted this question, but I felt that it was inappropriate for this site. I don't know where people buy used cars from doctors, but I suspect it is not on the forum of medical advisory. [cont] $\endgroup$ – Asaf Karagila♦ Nov 18 '12 at 18:21
  • $\begingroup$ [...] You should have looked around for the tags, maybe check the Tags page and see what's there. If you would have used [numerical-methods] it might have signaled that this is a question on a topic I am unfamiliar with, and I would have been more reluctant to close. $${}$$As for the last comment, this is merely a reply to your last comment which I found somewhat offensive. True no one left a comment, but when the site is getting bigger, and you close more questions, you stop leaving comments on all of them. You did well to come for the meta for this. [cont] $\endgroup$ – Asaf Karagila♦ Nov 18 '12 at 18:24
  • $\begingroup$ [...] As for the the fact that closing is overused, I disagree. Closing is an easily reversible action, in some cases the closing parties may vote to reopen as well. I agree that using Off-Topic may be off putting (pun intended) to the OP, but this is the way it is. The site is getting too big, and we cannot appease every single user. I still think that saying that high rep. users belittle others is grossly incorrect and quite offensive. I can vouch for myself that I spend almost 50% of my waking hours within "refresh distance" of the site, and I am sure I am not the only one. [cont] $\endgroup$ – Asaf Karagila♦ Nov 18 '12 at 18:26
  • $\begingroup$ [...] High reputation often correlates with being also active on meta (at least reading if not posting) and getting involved in voting, closing, deleting and reopening. People who do that do that for free. They are not paper pushing bureaucrats which just spend the day from 9 to 5. They spend their own free time, and when someone who is a newbie says something as you did makes me wonder who am I working so hard for, and who are the others working so hard for? Obviously our work goes unappreciated on the first sign of a possible mistake, even if that mistake is easily correctable. [cont] $\endgroup$ – Asaf Karagila♦ Nov 18 '12 at 18:28
  • $\begingroup$ [...] All in all, however, I am glad it worked for your benefit. I hope that in the next time you will try to make less accusations and focus on the positive parts. Remember that no one was out to get you, and nothing was done out of hard feelings. However things can differ, and there have been documented cases here where people that threw too many stones still get treated badly for that reason. I hope that this will be a lesson to anyone, that closing is not that horrible, and talking down to people who spend their time for free can become a real issue. $${}$$I think I'll stop now. :-) $\endgroup$ – Asaf Karagila♦ Nov 18 '12 at 18:31
  • $\begingroup$ @Asaf Karagila: I don't feel anything you wrote as rude (the "used cars from doctors" part made me laugh). I agree with you and thank you both for these comments and for your other work here. Forget about my "can't read" comment. I think I'll stop now, too. $\endgroup$ – maaartinus Nov 18 '12 at 20:23

You must log in to answer this question.

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