I am new to the Mathematics SE forum. I recently asked my first question. I received comments that it was a question about computing rather than maths. Today, I started receiving downvotes, which I suspect is for the same reason.

I strongly object because computational mathematics is a well recognised branch of mathematics, even if it is not necessarily well defined, i.e. it has considerable overlap with computing and other fields of maths. Moreover, my interest in this question comes from me as a mathematician, not a computer programmer/software developer/etc. I mention 1 particular software that I had attempted to solve this in, but am not seeking a solution that is language specific. In fact, I am not seeking code of any sort, but a general approach to this problem, such as a pseudo-code algorithm (see my answer to my own question).

I believe this stance is in accordance with the rules for Maths SE regarding what is on-topic. This is reiterated by the comments received by the OP of this similar similar question.

So, am right in asserting that computational maths questions are considered on-topic? Can this be made clearer in the help section, "what is on-topic"? Are there changes I can make to this question clearer so I stop receiving negative feedback?

Aside: I have a PhD in math, in a different field, not computational math. However, as an undergrad math student, I studied 2 units specifically on computational maths. I know first-hand that topics relating to machine precision are pertinent to the field of computational maths. When someone says, "your question isn't about math" either directly or by downvoting, it is actually quite insulting. As a community, I think we can do better.

  • 3
    $\begingroup$ Now this question asking for explanation is being downvoted? This community is feeling extremely toxic. $\endgroup$ – Earlien Jan 14 '20 at 9:05
  • 13
    $\begingroup$ First of all, voting on meta is about agreeing or disagreeing. Secondly, while computational mathematics is mathematics, asking how to store a big number in a programming language or another is certainly not computational mathematics or mathematics of any kind. Not more than "what is your preferred paper size for working mathematical ideas?" is a mathematical question. $\endgroup$ – Asaf Karagila Jan 14 '20 at 10:34
  • 1
    $\begingroup$ The question comes from a maths YoutTube channel (Numberphile), presented by a mathematician (Matt Parker), and is expanded by me, also a mathematician from a maths perspective. Working in finite precision IS fundamental to computational maths. There may be overlap, but I don't see your justification. $\endgroup$ – Earlien Jan 14 '20 at 10:55
  • 9
    $\begingroup$ I can cook you lunch, and then you can eat it and do the dishes. Just because the whole thing involved mathematicians doesn't make it mathematical. $\endgroup$ – Asaf Karagila Jan 14 '20 at 11:00
  • 3
    $\begingroup$ That analogy is absurdly hyperbolic and mischaracterises the complexity of the stated problem. $\endgroup$ – Earlien Jan 14 '20 at 11:04
  • 3
    $\begingroup$ I do not have strong views at the moment on whether the question is on or off topic, but I do wonder whether we are being sufficiently welcoming to a new contributor. $\endgroup$ – J W Jan 14 '20 at 13:55
  • 1
    $\begingroup$ Why don't you ask your question anyway on Stack Overflow. From the response you get there you will be able to judge which was the better platform for your question. $\endgroup$ – Peter Phipps Jan 14 '20 at 16:08
  • $\begingroup$ added computational msthematics tag to the original question ... $\endgroup$ – user645636 Jan 14 '20 at 16:57
  • $\begingroup$ computational-mathematics This tag concerns computational problems central to mathematical and scientific computing. The scope includes algorithms, numerical analysis, optimization, and linear algebra, computational topology, computational geometry, symbolic methods, and inverse problems $\endgroup$ – user645636 Jan 14 '20 at 17:11
  • $\begingroup$ @JW Could someone explain if it is off-topic, why is there a “machine precision” tag in maths SE? How is it not relevant when every aspect of the problem is clearly mathematical, albeit involving algorithms (I’ve acknowledged the overlap)? If this is so blatantly wrong, then why is the “what’s on topic” section so vague? $\endgroup$ – Earlien Jan 14 '20 at 20:53
  • $\begingroup$ Great. Now it’s been closed with “This question is not about mathematics, within the scope defined in the help center.” Not at all clear why. It appears entirely within scope to me. $\endgroup$ – Earlien Jan 14 '20 at 21:01
  • 1
    $\begingroup$ We have a dedicated thread for making the case for reopening a question, math.meta.stackexchange.com/questions/28692/… $\endgroup$ – Gerry Myerson Jan 15 '20 at 5:43
  • 1
    $\begingroup$ Thanks. But it would be helpful to know why it was closed in the first place (so I know what edits to make). My understanding is that I should try to edit the question first before resorting to that. $\endgroup$ – Earlien Jan 15 '20 at 6:05
  • $\begingroup$ I don't think it is necessarily helpful to point to the guidelines and say and tags and try and argue whether or not the question is on topic. This question was open for a week without a solid answer, and for this reason alone you should consider asking elsewhere. $\endgroup$ – user1729 Jan 15 '20 at 10:26
  • 4
    $\begingroup$ In summary: If you want an answer to your question then I suggest asking it on StackOverflow; not because of rules or guidelines, but simply because you are much more likely to get an answer there. $\endgroup$ – user1729 Jan 15 '20 at 11:10

Here's my rule of thumb for being on-topic:

Can this Question be resolved by reasoned mathematical argument?

Your post (on main Math.SE) ends with the following words:

So what am I doing wrong?

Do I need to store it as a float so that all the digits are right of the decimal point? Do I need a special library (like Brobdingnag in R) to handle values larger than double precision?

Your post's title Constant in Tupper's Self-Referential Formula too large for machine precision strongly suggests that you realize that double precision floating point representation is inadequate for the task at hand, and I see no reason to doubt that is precisely the case. However asking for "a special library" recommendation to provide multiple precision is indeed off-topic by my lights. It is not an issue that can be resolved by reasoned mathematical argument.

If you are indeed more on the mathematics side of things, then I can understand your interest in this topic. But there are many ways to handle the necessary extended precision, and it would be best for you to do your own research. Come back to Math.SE when you have a mathematical problem to solve, computational or otherwise, and I'll be happy to help as best I can.

  • $\begingroup$ Thanks for the input. I only offered the idea of a library to handle more precision as a suggestion - I still have no idea if that is feasible. But as you can see from the answer that I suggested, that was not the solution I proposed. As for doing my own research, I am the only one who offered an answer. I still consider it a mathematical problem, and cannot see how this is off-topic according to the rules (which are broader than your rule of thumb). $\endgroup$ – Earlien Jan 14 '20 at 21:58
  • 6
    $\begingroup$ The provision of extended precision arithmetic is exigent upon your computing platform. It is software issue. If you are asking for software recommendations, this isn't the right forum. I take it that you see that without the extended precision, that constant cannot be exactly represented, and the effects of using a floating point approximation are platform dependent. $\endgroup$ – hardmath Jan 14 '20 at 22:08
  • $\begingroup$ Yes I understand that would indeed be a software issue. The only reason I mentioned a specific package was to show what I had tried. I certainly did not ask for a solution in a specific language. Nor did I specifically ask for a way of implementing in extended precision - it was just an idea I floated (pun intended). I was open to any solution. $\endgroup$ – Earlien Jan 14 '20 at 22:18
  • $\begingroup$ Does my answer below address your last sentence i.e. Is it clearer now that this is a mathematical problem? $\endgroup$ – Earlien Jan 15 '20 at 0:23
  • 9
    $\begingroup$ @Earlien: It did not. I went back through your original post again (more than once) and tried to identify what mathematical problem you are raising. Setting aside the interrogative sentences which you say merely suggestions for solutions, and not the essential problem statement, I'm at a loss to find much specific in your post that does not say, How can I represent a 543 digit number on a computer? If you intend to revise the Question, I'd encourage you to consider what mathematical problem would exist without reference to computer technology. $\endgroup$ – hardmath Jan 15 '20 at 22:19

You must log in to answer this question.

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