I don't know if it should matter but the sum of the intercepts the lines make with x and y axis is constant I think. It looks like a hyperbola to me.
-
1Hello guitar man, are we missing some picture here? – imranfat Mar 06 '16 at 17:39
-
I'm on mobile. Added the link in the description – user320390 Mar 06 '16 at 17:48
-
I don't have enough reputation to post an image:( – user320390 Mar 06 '16 at 17:56
-
That picture looks nice... – imranfat Mar 06 '16 at 18:04
-
Similar question asked earlier. I guess it's just in the air! – pjs36 Mar 06 '16 at 18:05
2 Answers
We are given a family of lines $$\ell_c:\quad{x\over1+c}+{y\over 1-c}=1\qquad(-1<c<1)\ .$$ A given point $(x,y)$ belongs to one of the lines iff the equation $$x(1-c)+y(1+c)=1-c^2\tag{1}$$ has real solutions $c$. The boundary between the covered and the uncovered part of the plane is given by the points $(x,y)$ for which the discriminant of the quadratic equation $(1)$ vanishes. This amounts to the equation $$(y-x)^2-4(x+y-1)=0\ .\tag{2}$$ After introducing new orthonormal coordinates $$u:={1\over\sqrt{2}}(y-x),\qquad v:={1\over\sqrt{2}}(y+x)$$ equation $(2)$ assumes the form $v=p u^2 +q$ with certain constants $p$ and $q$. It follows that the locus in question is an arc of a parabola.
- 226,825
Parabola if you are joining equispaced divisions by a straight line.
EDIT 1:
The notches or holes are assumed equi-spaced. Let the segments be $ a-u, a+u $ along y and x-axes.
$$ \frac{x}{a-u} + \frac{x}{a+u}= 1 \tag{1}$$
To find envelope by C-discriminant method to find envelope.
$$ \frac{x}{(a-u)^2} - \frac{x}{(a+u) ^2} = 1 \tag{2} $$
$$ \frac{\sqrt x}{\sqrt y }= \frac{a-u}{a+u} \tag{3} $$
$$ \frac{u}{a} = \frac{{\sqrt y}-{\sqrt x} }{{\sqrt y}+{\sqrt x} }\tag{4} $$
Plug (1) into (4) and simplify to get a parabola envelope tangent to axes:
$$ {\sqrt x}+{\sqrt y}= \sqrt { 2 a}. \tag{5} $$
which is a parabola with axes at $45^0$ to x,y axes and tangent to them.
- 40,495
-
The OP wrote "It looks like a hyperbola to me." I think in the circumstances your brief assertion(?) that this is instead a parabola needs justification. – hardmath Mar 06 '16 at 19:03
-
A while ago I answered on "loci of intersection..." in a more detailed manner. To my judgment the product of x,y coordinates is not constant. To me it appears the sum of the intercepts is constant.I give another justification/ proof in the edit for same result as parabola. Thanks – Narasimham Mar 06 '16 at 19:39
