Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Discussion Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to nForum
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorKeithEPeterson
    • CommentTimeDec 8th 2016
    • (edited Dec 8th 2016)

    Yesterday I defined a function that replaces decimal digits equal to yy of the number xx with the arbitrary number zz. Assuming I got it right: sgn(x) n=log 10|x| y=10 n1|x|1010 n2|x|z+ y10 n1|x|1010 n2|x|10 n1|x|1010 n2|x|10 nsgn(x)\cdot\sum_{n=-\left\lfloor log_{10} |x| \right\rfloor}\frac{\sum_{y=\left \lfloor 10^{n-1}|x|\right\rfloor-10\left\lfloor10^{n-2}|x|\right\rfloor}z+\sum_{y\ne\left \lfloor 10^{n-1}|x|\right\rfloor-10\left\lfloor10^{n-2}|x|\right\rfloor}\left \lfloor 10^{n-1}|x|\right\rfloor-10\left\lfloor10^{n-2}|x|\right\rfloor}{10^{n}}

    Maybe I could implement the lambda calculus over the Reals for giggles…

    Have any of you partaken in recreational math? If yes, what kind of stuff were you working on?

    • CommentRowNumber2.
    • CommentAuthorTodd_Trimble
    • CommentTimeDec 8th 2016

    Well, yes! For example there are problem sections in journals like the American Mathematical Monthly, and I often take a crack at those, or at Putnam type problems, etc. etc. Sometimes I’ll work on silly things just for kicks, like characterizing the triangular squares or evaluating tricky integrals. Generally speaking I have a taste for generatingfunctionology, binomial identities, finite difference calculus…

    Answering questions on MO or Mathematics StackExchange can often have a recreational feel.

    To be honest, I don’t make a huge distinction between “recreational math” and “serious math”. For me both involve a certain amount of playing and jiggering around, to see what works. Concrete questions and abstract considerations are frequently mixed together; it’s a big potpourri. This is very true in number theory for example, and I do a certain amount of diddling around, writing things up for my own amusement (recently it was cooking up a home-made proof of Lagrange’s four square theorem, avoiding the usual pigeonhole arguments).

    One of my first loves in mathematics was continued fractions, and the old stuff you might find in HAKMEM has a certain appeal to me. I would like to understand better some of Gosper’s algorithms for doing arithmetic of continued fractions from an algebraic/coalgebraic perspective – that’s the type of thing I mean when I say that abstract and the concrete are all mixed together. But I’m hardly alone in this. It’s pretty clear to me that a lot of category theorists love this type of fooling around. The late Steve Schanuel is an excellent example, as is Peter Freyd, and probably Michael Barr, and probably Andreas Blass, to name a few examples just off the top of my head.

    • CommentRowNumber3.
    • CommentAuthorKeithEPeterson
    • CommentTimeDec 8th 2016
    • (edited Dec 8th 2016)

    Wouldn’t a continued fraction be the cardinality class of \infinity-groupoids? Consider the groupoids with the cardinality of the golden ratio for instance. One would think they should have a beautiful self-recursive interpretation if given a topology.

    Edit: Yeah. The cardinality of such a groupoid is given by the hairy expression, |ϕ|= isoclassesofobjects[ϕ]1+1 isoclassesofobjects[ϕ]1+1 isoclassesofobjects[ϕ]1+1 isoclassesofobjects[ϕ]1+1|\phi| = \sum_{iso classes of objects[\phi]}1+\frac{1}{\sum_{iso classes of objects[\phi]}1+\frac{1}{\sum_{iso classes of objects [\phi]}1+\frac{1}{\sum_{iso classes of objects[\phi]}1+\frac{1}{\ddots}}}},

    showing that there is one such groupoid, and what this thing really is: A groupoid that has a disjoint object next to a groupoid, that has a disjoint object next to a groupoid,…

    Who says math can’t be fun?

    • CommentRowNumber4.
    • CommentAuthorKeithEPeterson
    • CommentTimeDec 8th 2016

    Actually, the page continued fraction on nlab doesn’t mention groupoid cardinality. I feel it should.

    • CommentRowNumber5.
    • CommentAuthorMike Shulman
    • CommentTimeDec 8th 2016

    I certainly never do any math that I don’t find fun!

    • CommentRowNumber6.
    • CommentAuthorKeithEPeterson
    • CommentTimeJan 1st 2017
    • (edited Jan 2nd 2017)

    Oh, now here’s a nice continued fraction expression for π\pi,

    π=2+2+2+2+1+21+21+21+21+21+21+21+21+21+21+21+21+21+21+2, \pi =2+\frac{2+\frac{2+\frac{2+\frac{\vdots }{1+\frac{2}{1+\frac{2}{\ddots }}}}{1+\frac{2}{1+\frac{2}{1+\frac{2}{\ddots }}}}}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{\ddots }}}}}}}{1+\frac{2}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{1+\frac{2}{ \ddots }}}}}}}} ,

    it follows from Viète’s formula and Engel expansion.

    Edit: math typo in the expression, thank you Todd for pointing it out.

    • CommentRowNumber7.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 2nd 2017

    That doesn’t look right to me. Isn’t this claiming π=2+π2\pi = 2 + \frac{\pi}{\sqrt{2}}?

    • CommentRowNumber8.
    • CommentAuthorKeithEPeterson
    • CommentTimeJan 2nd 2017
    • (edited Jan 2nd 2017)

    If,abcdefgh=d+c+b+aefgh\frac{a b c d}{e f g h} = \frac{d+\frac{c+\frac{b+\frac{a}{e}}{f}}{g}}{h},

    and if π2=2222+222+2+2\frac \pi2 = \frac 2{\sqrt 2} \cdot \frac 2{\sqrt{2+\sqrt 2}} \cdot \frac2{\sqrt{2+\sqrt{2+\sqrt 2}}} \cdots,

    then π=(2222+222+2+2)21\pi = ( \frac 2{\sqrt 2} \cdot \frac 2{\sqrt{2+\sqrt 2}} \cdot \frac2{\sqrt{2+\sqrt{2+\sqrt 2}}} \cdots)\cdot\frac2 1,

    with, a+a+a+a+=1+a1+a1+a\sqrt{a+ \sqrt{a+ \sqrt{a+ \sqrt{a+ \cdots }}}} = 1+\frac{a}{1+\frac{a}{1+\frac{a}{}\ddots }},

    then, π=2+2+2+2+1+21+21+21+21+21+21+21+21+21+21+21+21+21+21+2,\pi =2+\frac{2+\frac{2+\frac{2+\frac{\vdots }{1+\frac{2}{1+\frac{2}{\ddots }}}}{1+\frac{2}{1+\frac{2}{1+\frac{2}{\ddots }}}}}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{\ddots }}}}}}}{1+\frac{2}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{1+\frac{2}{ \ddots }}}}}}}} ,

    Edit: Sorting out fraction.

    • CommentRowNumber9.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 2nd 2017
    • (edited Jan 2nd 2017)

    [never mind]

    • CommentRowNumber10.
    • CommentAuthorKeithEPeterson
    • CommentTimeJan 2nd 2017
    • (edited Jan 2nd 2017)

    You are right by the way, by switching the values 1 and 2 in the infinite expansion, it became 2\sqrt2. Which is baaaaad.

    • CommentRowNumber11.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 2nd 2017

    π=2+2+2+2+1+21+21+21+21+21+21+21+21+21+21+21+21+21+21+2,\pi =2+\frac{2+\frac{2+\frac{2+\frac{\vdots }{1+\frac{2}{1+\frac{2}{\ddots }}}}{1+\frac{2}{1+\frac{2}{1+\frac{2}{\ddots }}}}}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{\ddots }}}}}}}{1+\frac{2}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{1+\frac{2}{ \ddots }}}}}}}} ,

    I believe

    x=1+21+21+21+21+21+2x = 1+\frac{2}{1+\frac{2}{1+\frac{2}{{1+\frac{2}{1+\frac{2}{1+\frac{2}{ \ddots }}}}}}}

    reduces to x=2x = 2, since it satisfies x=1+2xx = 1 + \frac{2}{x}. As a result, it sure looks to me that the cited expression for π\pi (let’s denote it yy) satisfies the equation y=2+y2y = 2 + \frac{y}{2}.

    • CommentRowNumber12.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 2nd 2017

    While we’re on the topic, let me recall that continued fractions which exhibit the type of regularity seen in some of your expressions are fixed points of fractional linear transformations over \mathbb{Z}, i.e. satisfy x=ax+bcx+dx = \frac{a x + b}{c x + d} where a,b,c,d,adbc=±1a, b, c, d \in \mathbb{Z}, a d - b c = \pm 1. Such xx thus belong to quadratic number fields. Therefore it might be tricky getting nice expressions for e.g. 2+2\sqrt{2 +\sqrt{2}} which aren’t quadratic surds. (So in general, check your work carefully.)

    • CommentRowNumber13.
    • CommentAuthorKeithEPeterson
    • CommentTimeJan 2nd 2017

    Wow, I defined the value 2.5, not π\pi. Bravo me. Not so fun with math.

    But mistakes are part of math.

    • CommentRowNumber14.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 2nd 2017

    Well, I got y=4y = 4, not y=2.5y = 2.5. ;-)

    • CommentRowNumber15.
    • CommentAuthorKeithEPeterson
    • CommentTimeJan 2nd 2017
    • (edited Jan 2nd 2017)

    Ugh, I hate ascending fractions!!!

    • CommentRowNumber16.
    • CommentAuthorKeithEPeterson
    • CommentTimeJan 2nd 2017
    • (edited Jan 2nd 2017)

    This is just proof that counting down from infinities is a bad idea.

    And let me be clear, this is a really bad attempt at a hack.

    Edit: No one makes mistakes in math, right? I mean, especially not in front of a classroom. And especially especially not when sending in work to be refereed.

    • CommentRowNumber17.
    • CommentAuthorTodd_Trimble
    • CommentTimeJan 2nd 2017
    • (edited Jan 2nd 2017)

    Re #15: I think you probably want b+b 2+4a2\frac{b + \sqrt{b^2 + 4a}}{2}, right? And here, we should get 1+1+82=2\frac{1 + \sqrt{1 + 8}}{2} = 2, as I had said before.

    (Since my comments, such as this, seem to be made irrelevant by repeated edits/erasures by Keith, I think what I shall do is not comment for a while on his posts.)