Skip to main content
38 events
when toggle format what by license comment
Nov 15, 2019 at 15:19 comment added Kapil Any proof in number theory that involves real numbers uses them to estimate a certain interval where some number lies. This estimate can always be weakened a bit to use only rational numbers. In fact, one could give a course in "Real Analysis" that uses only intervals and such intervals could be limited to those with rational boundaries. Thus, with a lot of effort, real numbers could be eliminated from the discussion entirely. Is it worth the effort? Perhaps, if one wants to make things "computationally effective".
Nov 15, 2019 at 9:44 history made wiki Post Made Community Wiki by Todd Trimble
Nov 15, 2019 at 1:35 answer added Olivier timeline score: 1
Nov 14, 2019 at 22:10 answer added Timothy Chow timeline score: 4
Jan 11, 2012 at 2:14 vote accept Daniel Moskovich
Apr 17, 2011 at 23:29 answer added John Stillwell timeline score: 10
Apr 17, 2011 at 16:00 history edited Daniel Moskovich CC BY-SA 3.0
added 837 characters in body
Apr 15, 2011 at 12:01 comment added Daniel Moskovich @Lubin: I'm arguing not that mathematics should be fragmented; rather that everything we do in mathematics should have a clearly defined/understood purpose, including the choice of tools used in a proof.
Apr 15, 2011 at 2:26 comment added Ramsey Apropos the "utopian" comment/debate, I'm finding myself of two minds that I cannot seem to disentangle. As an example relevant to this thread, I'd have to say that I'm fine with the idea of solving a global problem by local means (eg looking at congruence to solve diophantine problems) but somehow less comfortable with the opposite (eg the Harris-Taylor proof of local langands). I really can't exactly characterize why. Perhaps a measure of canonicality? Or maybe "upstructuring" vs "downstructuring"? Can anyone help diagnose my affliction?
Apr 15, 2011 at 1:02 comment added Ryan Budney Related thread: math.stackexchange.com/questions/5303/…
Apr 14, 2011 at 23:24 comment added user9072 @Daniel Moskovich: I tried to give a selection of examples, some noninteresting to make a point and some perhaps interesting. Perhaps, I managed to include something you like; in any case, I found the task of doing so interesting. So thanks for the question.
Apr 14, 2011 at 22:43 comment added Daniel Moskovich @unknown(google): ... where XYZ is "the dual/associated/whatever something isn't an integer, and it would be artificial to force it to be so". That would be one thing which would make me happy.
Apr 14, 2011 at 22:42 comment added Daniel Moskovich @unknown(google): Your answer is excellent! I'm just a bit slow... Regarding your request, I don't know if this is the best example, but let's say Poincare duality in the simplicial world is most easily understood by passing to the more general setting of CW complexes, because the dual cell structure isn't a triangulation, and to triangulate would be artificial. So a "best" proof of Poincare duality for simplicial complexes would pass through CW complexes. I want to see something analogous for integers, in a clear simple context- it's most natural to adjoin reals because XYZ
Apr 14, 2011 at 22:11 comment added user9072 May I ask that you provide some example of what you envision in any context whatsoever. That is an example of an application of (truly, so excluding what you called formal constants) real numbers wherever you like in which one can not more or less trivially poke holes by saying 'Look just successively approximate this. That it will be a total mess then, I don't mind.' That being said I will try one more update of my answer.
Apr 14, 2011 at 21:22 history edited Daniel Moskovich CC BY-SA 3.0
added 1062 characters in body
Apr 14, 2011 at 19:38 answer added Charles timeline score: 9
Apr 14, 2011 at 19:33 answer added anonymous timeline score: 12
Apr 14, 2011 at 16:51 answer added François Brunault timeline score: 8
Apr 14, 2011 at 15:17 answer added stankewicz timeline score: 4
Apr 14, 2011 at 14:42 answer added GH from MO timeline score: 18
Apr 14, 2011 at 12:20 answer added BSteinhurst timeline score: 3
Apr 14, 2011 at 10:56 comment added user9072 @Quanta: what do you mean by 'in a finite way' precisely? Anyway, let $s(x)$ denote the number of natural numbers up to size $x$ not divisible by the square of a prime number (squaree free numbers). Then $s(x)/x$ converges to $6/\pi^2$. Also see Gerry Myerson's answer.
Apr 14, 2011 at 10:05 comment added Quanta Can you please give some examples of real numbers being useful in number theory? Hopefully not algebraic numbers because they can be constructed in a finite way unlike the other trancendental real numbers..
Apr 14, 2011 at 9:39 answer added François G. Dorais timeline score: 80
Apr 14, 2011 at 9:37 answer added user9072 timeline score: 8
Apr 14, 2011 at 8:33 comment added S. Carnahan What you suggest would be a utopia for people who don't want to learn new tools. We seem to have strong experimental evidence that fluency in one area is not a good way to make leaps in our ability to prove or understand mathematics. The question is somewhat related to the problem of whether tersely stated theorems should have short proofs - if the length of proofs were bounded by a computable function in the length of the theorem, we could check any mathematical statement in finite time.
Apr 14, 2011 at 8:14 answer added Sidney Raffer timeline score: 19
Apr 14, 2011 at 5:26 comment added Lubin The real moral of the story is that there aren't separate ``worlds'' in mathematics, but rather a beautiful and often mysterious all-embracing unity.
Apr 14, 2011 at 5:15 answer added Ramsey timeline score: 6
Apr 14, 2011 at 4:08 answer added Gerry Myerson timeline score: 8
Apr 14, 2011 at 3:30 answer added Faisal timeline score: 27
Apr 14, 2011 at 3:29 comment added Daniel Moskovich @Mariano: The hypothetical argument states possible existence of "correctly tooled" arguments, even ones which aren't known and won't be known for the next million years.
Apr 14, 2011 at 3:13 answer added Qiaochu Yuan timeline score: 8
Apr 14, 2011 at 3:09 comment added Mariano Suárez-Álvarez Well, the hypothetical argument is pretty much against reality... doesn't that refute it by itself? One can also come up with «maybe partial differential equations are the wrong tool for the Poincaré conjecture, and maybe someone will come up with a purely topological argument and win a Fields^2 medal for it», and so on. The thing is, such alternative, correctly tooled arguments are slow in the coming!
Apr 14, 2011 at 3:05 comment added Qiaochu Yuan I think Alon is disagreeing with the sentiment of the first clause of the first sentence rather than the entire first sentence.
Apr 14, 2011 at 2:50 comment added Daniel Moskovich @Alon: "unless there were a good conceptual reason for the contrary". Are you arguing that statements and proofs should not live in the same world, and moreover, that the reasons for this should be unclear?
Apr 14, 2011 at 2:45 comment added Alon Amit This is totally tangential but I just want to say I strongly disagree with the sentiment of the first sentence. If statements and proofs were typically "in the same world", math would be so dull. Dystopia.
Apr 14, 2011 at 2:40 history asked Daniel Moskovich CC BY-SA 3.0