| Main index | Other Papers index | About author |
Julian D. A. Wiseman
Contents: if the boundary of an n/m star is drawn, there are n points on the exterior bounding circle, and another n points on an interior circle (see diagram). The table below shows, in surds, the radius of this interior circle, for several values of n and m: n/m general case, 5/2, 6/2, 8/2, 8/3, 10/2, 10/3, 10/4, 12/2, 12/3, 12/4, 12/5, 15/2, 15/3, 15/4, 15/5, 15/6, 15/7, 16/2, 16/3, 16/4, 16/5, 16/6, 16/7, 20/2, 20/3, 20/4, 20/5, 20/6, 20/7, 20/8, and 20/9.
Publication history: only here. Usual disclaimer and copyright terms apply. Also see the values of Sin[] and Cos[] in surds, and the values of Cosecant[] = Cosec[] = Csc[] = 1/Sin[] and Secant[] = Sec[] = 1/Cos[] in surds.
![]() |
If the boundary of an n/m star is drawn, there are n points on the exterior bounding circle, and another n points on an interior circle (see diagram on right). The table below shows, in surds, the radius of this interior circle, for several values of n and m. The surds are are shown in several formats.
Graphical formula: a .png, derived from…
LaTex: a LaTeX expression.
Excel: copy-pasteable into Excel, which will automatically convert the “Sqrt” into an upper-case “SQRT”.
CalcCenter: if one enters Sqrt[2] directly into Mathematica CalcCenter (the budget version of Mathematica), it automatically evaluates the expression numerically, frustrating an attempt to work with surds. To prevent this integers inside the inner-most Sqrts have been replaced with the likes of Int5, which CalcCenter treats as a variable. If using the full-expense Mathematica instances of “Int” may be removed by a preprocessor or by setting Int2=2; Int3=3; Int5=5;.
Postscript: using 3 5 sqrt sub 2 div rather than 180 5 div dup 2 mul cos exch cos div is about as efficient computationally, but is, in some sense, far more elegant.
Help! Endeavours have been made to represent these values as simply as possible. But further simplifications would be welcomed, credit being given.
Trivia: as m→+∞ with i fixed, the inner radius of the (2m+i)/m star tends downwards to i/(2+i) + m–2×π2i(i+1)/(24(i+2)) + O(m–4), and hence the inner radii are bounded below by ⅓.
Open question: stars (6i–2)/i and (18i–6)/(6i–2) have the same inner radius, as do (6i–4)/i and (18i–12)/(6i–3), for integer i≥2 (PDF of overlapping 8/2 and 24/9 stars, and of 10/2 and 30/10 and 14/3 and 42/15 stars). This is easily shown using the double-angle formulae. But are all pairs of stars with matching inner radii in one of these two forms? A proof (or, less probably, a counterexample) would be welcomed. (Also observe that the inner radius of both forms tends to unity, being >0.99 when i≥31, and >0.99975 when i≥1210.)
Errors: whilst the outputs have been tested, it is possible that errors remain. Please do test things before embedding them somewhere important—and if errors or possible improvements are found, tell the author.
Julian D. A. Wiseman, June 2008
| Main index | Top | About author |