x1.Fx' (x,y,z) + y1.Fy' (x,y,z) + z1.Fz' (x,y,z) = 0 <=> x.Fx' (x1,y1,z1) + y.Fy' (x1,y1,z1) + z.Fz' (x1,y1,z1) = 0 with matrix notation: [Fx' (x1,y1,z1)] <=> [x y z ] [Fy' (x1,y1,z1)] = 0 [Fz' (x1,y1,z1)] <=> PT C P1 = 0 <=> P1T C P = 0
Proof:
A(x1,y1,z1) is not on conic section F(x,y,z) = 0.
Say B(x,y,z) is an arbitrary point.
A variable point P of the line AB has coordinates
(x + h x1, y + h y1, z + h z1) Point P is on the conic section <=> F(x + h x1, y + h y1, z + h z1) = 0 <=> F(x,y,z) + h (x.Fx' (x1,y1,z1) + y.Fy' (x1,y1,z1) + z.Fz' (x1,y1,z1)) + h2 F(x1,y1,z1) = 0This is a quadratic equation in h. The roots h1 and h2 correspond with the intersection points of the conic section and AB.
Points A and B are harmonic conjugate points with respect to the intersection points P1 and P2 of the conic section and the variable line AB
<=> (P1,P2,A,B) = -1 <=> (A,B,P1,P2) = -1 <=> h1 = - h2 <=> h1 + h2 = 0 <=> x.Fx' (x1,y1,z1) + y.Fy' (x1,y1,z1) + z.Fz' (x1,y1,z1) = 0 <=> B is element of the stated setWe see that the equation of the polar line and the equation of the stated set are the same.
If that point A(x1,y1,z1) is a simple point then:
A(x1,y1,z1) is on its polar line <=> x1.Fx' (x1,y1,z1) + y1.Fy' (x1,y1,z1) + z1.Fz' (x1,y1,z1) = 0 <=> 2.F(x,y,z) = 0 <=> point A(x1,y1,z1) is on the conic section
Proof:
Take a point A(x1,y1,z1), different from a double point.
The polar line of A is x1.Fx' (x,y,z) + y1.Fy' (x,y,z) + z1.Fz' (x,y,z) = 0
Say D(xo,yo,zo) is a double point.
We investigate if D is on the polar line.
x1.Fx' (xo,yo,zo) + y1.Fy' (xo,yo,zo) + z1.Fz' (xo,yo,zo) = x1 . 0 + y1 . 0 + z1 . 0 = 0Thus, D is on the polar line.
A(x1,y1,z1) is on the polar line of B(x2,y2,z2) <=> x1.Fx' (x2,y2,z2) + y1.Fy' (x2,y2,z2) + z1.Fz' (x2,y2,z2) = 0 <=> x2.Fx' (x1,y1,z1) + y2.Fy' (x1,y1,z1) + z2.Fz' (x1,y1,z1) = 0 <=> B(x2,y2,z2) is on the polar line of A(x1,y1,z1)
Point A is a pole of line a with respect to a conic section <=> Line a is a polar line of A.
The conic section has equation PT C P = 0 The line a has equation u x + v y + w z = 0 <=> [x] [u v w] [y] = 0 [z] <=> U.P = 0 with U = [u v w] [x1] Say P1 = [y1] are the coordinates of a possible pole of line a. [z1] The polar line is P1T C P = 0 This polar line is also the line U.P = 0 From this we have : k U = P1T C <=> P1T = k U C-1 From that last result, we see that there is exactly one pole of the line a.The last formula gives a method to calculate the pole.
Take the conic section y2 - 2 x = 0 and the line x + y + 1 = 0We'll show three methods to calculate the pole of this line.
[ 0, 0, -1 ] C = [ 0, 1, 0 ] U = [1 1 1] [ -1, 0, 0 ] -1 [ 0, 0, -1 ] C = [ 0, 1, 0 ] [ -1, 0, 0 ] T [ 0, 0, -1 ] P1 = k [1 1 1]. [ 0, 1, 0 ] = k [-1 1 -1] [ -1, 0, 0 ] The pole is point (-1,1,-1) or (1,-1,1)
x1.Fx' (x,y,z) + y1.Fy' (x,y,z) + z1.Fz' (x,y,z) = 0 <=> x1(-2z) + y1(2y) + z1(-2x) = 0 <=> z1 x - y1 y + x1 z = 0This line has to be the line x + y + z = 0. So the pole is point (1,-1,1).
The points A and B are conjugated with respect to a conic section <=> A is on a polar line of point B <=> B is on a polar line of point A.Remark :
A(x1,y1,z1) and B(x2,y2,z2) are conjugated points <=> A is on a polar line of point B <=> x1.Fx' (x2,y2,z2) + y1.Fy' (x2,y2,z2) + z1.Fz' (x2,y2,z2) = 0
A(x1,y1,z1) and B(x2,y2,z2) are conjugated points <=> B is an arbitrary point <=> x2.0 + y2.0 + z2.0 = 0 <=> x2.Fx' (x1,y1,z1) + y2.Fy' (x1,y1,z1) + z2.Fz' (x1,y1,z1) = 0 <=> x1.Fx' (x2,y2,z2) + y1.Fy' (x2,y2,z2) + z1.Fz' (x2,y2,z2) = 0
A(x1,y1,z1) and B(x2,y2,z2) are conjugated points <=> x2.Fx' (x1,y1,z1) + y2.Fy' (x1,y1,z1) + z2.Fz' (x1,y1,z1) = 0 <=> x1.Fx' (x2,y2,z2) + y1.Fy' (x2,y2,z2) + z1.Fz' (x2,y2,z2) = 0
The lines a and b are conjugated with respect to a conic section <=> a contains each pole of line b AND b contains each pole of line aRemark :
The lines a and b are conjugated with respect to a conic section <=> a contains the pole of b <=> b contains the pole of a
Remark :
Each ordered quartet of concurrent lines a,b,c,d is transformed in
an ordered quartet of collinear points A,B,C,D.
Each ordered quartet of concurrent points A,B,C,D is transformed in
an ordered quartet of collinear lines a,b,c,d .
A(x1,y1,z1) B(x2,y2,z2) C(x1 + h x2, y1 + h y2, z1 + h z2) D(x1 + h'x2, y1 + h'y2, z1 + h'z2)Then the cross ratio (A,B,C,D) = h/h'.
a( Fx' (x1,y1,z1) , Fy' (x1,y1,z1) , Fz' (x1,y1,z1) ) b( Fx' (x2,y2,z2) , Fy' (x2,y2,z2) , Fz' (x2,y2,z2) ) c( Fx' (x1,y1,z1) + h Fx' (x2,y2,z2) , ..., ...) d( Fx' (x1,y1,z1) + h' Fx' (x2,y2,z2) , ..., ...)The cross ratio (a,b,c,d) is h/h' .
We say that the polar triangle is conjugated to the conic section or that the conic section is conjugated to the polar triangle.
k A2 + l B2 + m C2 = 0k,l and m are homogeneous parameters (not all 0).
Proof:
B = 0 is polar line of point B => (B,C,S1,S2) = -1 => (AB,AC,AS1,AS2) = -1 => There is a value of h such that Line AS1 has equation B + h C = 0 and Line AS2 has equation B - h C = 0Now we take the system of conic sections with basic conic sections:
(B + h C).(B - h C) + k A.A = 0 <=> B2 - h2 C2 + k A2 = 0
k A2 + l B2 + m C2 = 0is conjugated to a triangle with A=0 ; B=0 and C=0 as equations of the sides.
Proof:
Since the conic section is non-degenerated, the parameter l is not 0.
Dividing the equation by l, the conic section has equation
B2 + (m/l) C2 + (k/l) A2 = 0 2 We denote m/l as - h <=> B2 - h2 C2 + (k/l) A2 = 0 <=> (B - h C) (B + h C) + (k/l) A2 = 0This is an element of the system of conic sections with basic conic sections:
k A2 + l B2 + m C2 = 0k,l and m are homogeneous parameters (not all 0).
Proof:.. Say the conic section is degenerated in the lines d1 and d2.
C = 0 is polar line of poin C. => (B,C,S1,S2) = -1 => (AB,AC,AS1,AS2) = -1 => There is a value of h such that Line AS1 has equation B + h C = 0 and Line AS2 has equation B - h C = 0 => The degenerated conic section has equation B2 - h2 C2 = 0
A2 = 0 or B2 = 0 or C2 = 0
k A2 + l B2 + m C2 = 0is conjugated to a triangle with A=0 ; B=0 and C=0 as equations of the sides.
Proof:
B2 - h2 C2 = 0 <=> (B - h C) (B + h C) = 0Then the conic section is conjugated to the triangle.