1

I am aware that the following expression represents the two angular bisectors for two straight lines.

$$\frac{a_{1}x+b_{1}y-c_{1}}{\sqrt{a_{1}^{2}+b_{1}^{2}}}=\pm \frac{a_{2}x+b_{2}y-c_{2}}{\sqrt{a_{2}^{2}+b_{2}^{2}}}\qquad $$

I had the following question which was already asked and answered here:

Is there any link between the sign of RHS and the bisector of the smallest (biggest) angle?

I then had the following questions regarding the best answer:

Forget about $c_1$ and $c_2$, put $u:=(a_1,b_1)/\sqrt{a_1^2+b_1^2}$, $v:=(a_2,b_2)/\sqrt{a_2^2+b_2^2}$ and let $z:=(x,y)$. The lines $u\cdot z=0$ and $v\cdot z =0$ are parallel to your lines $g_1$ and $g_2$. If $u\cdot v>0$ (i.e., $u$ and $v$ enclose an acute angle) then it easy to see that $u+v$ is orthogonal to the bisector of the smaller angle between $g_1$ and $g_2$, so this bisector is parallel to the line $(u+v)\cdot z=0$. If, on the other hand, $u\cdot v<0$ then $u$ and $-v$ enclose an acute angle; therefore $u-v$ is orthogonal to the bisector of the smaller angle between $g_1$ and $g_2$, so in this case the desired bisector is parallel to the line $(u-v)\cdot z=0$.

  1. What are $u$ and $v$ and $z$? Hence what would $u+v$ and $u-v$ be? (I am unable to understand the notation.)
  2. How are the dot products parallel to the original lines? (Probably answerable from 1.)
  3. Is there a way to use the original equations to tell if a bisector is 'internal' or 'external'?

Apologies for the silly question and for violating any rules. I do not have enough reputation to ask in the comments.

MarianD
  • 2,953
idunno
  • 75
  • 10
  • Re: 3: the sign of $u\cdot v$ is equal to the sign of $(a_1,b_1)\cdot(a_2,b_2)$. – amd Mar 12 '19 at 06:05

1 Answers1

1

Ad $1$:

$u, v$ are unit vectors perpendicular to those lines. $z := (x, y)$ is a $formal$ vector, so for example

$$u \cdot z = 0 $$

is an equation of a line, because it is

$$(u_1, u_2) \cdot(x,y) = 0$$ or $$u_1x+u_2y=0$$ $(u+v)$ and $(u-v)$ are perpendicular to bisectors $f, g$ of original lines $m,n$:

enter image description here

Ad $2$:

As you guessed, it follows from the previous part.

Ad $3$:

Yes, there is. Follow the link in your own question to learn how.

MarianD
  • 2,953
  • The diagram does help tremendously as I am trying to work my head around 1 and 2. But how does Ad.3 tell us which equation refers to the 'internal' bisector and which to the 'external'? So when '$+$' is used you get one of the bisectors and the other is obtained when '$-$' is used, but which one bisects the acute and which one the obtuse? – idunno Mar 11 '19 at 13:03
  • It doesn't tell us, I showed that it is impossible. The problem is that the same line has (infinitely) many different equations. – MarianD Mar 11 '19 at 13:10
  • Regarding 1 and 2. : Since $u,v$ are perpendicular to $m,n$ thus the vector $u+v$ (which is the acute bisector for $u,v$) is the 'obtuse' bisector for the original pair of lines. Thus $(u+v)\cdot z=0$ which is perpendicular to $u+v$ is the 'acute bisector' for the original lines?! Sorry for being painfully slow but why is $u\cdot z = 0$ parallel to the original? – idunno Mar 11 '19 at 13:29
  • 1
    See, you put a question and I answered it. You didn't accept my answer. It seems that you are not satisfied with it, so you have wait for a better one. – MarianD Mar 11 '19 at 15:04
  • Sorry for being a bit dense but can I ask follow up questions? – idunno Mar 11 '19 at 16:44
  • You didn't accept my answer (you didn't click on check mark near it). So I suppose that you are not satisfied with it. – MarianD Mar 11 '19 at 17:19
  • You’re mistaken Ad 3. The sign of $(a_1,b_1)\cdot(a_2,b_2)$, which is the same as the sign of $u\cdot v$, determines which of the two equations describes the acute bisector. (It’s the one with the opposite sign.) If you negate the equation of one of the lines, it does negate the sign in the bisector equation as you’ve noted, but it also negates the sign of this dot product, so the rule still holds. See https://math.stackexchange.com/q/2403530/265466. – amd Mar 12 '19 at 06:15
  • @amd - Oh, you are right! I mistakenly answered the quoted original question: "Is there any link between the sign of RHS and the bisector of the smallest (biggest) angle?" Thanks for it - I'm going to correct my answer. – MarianD Mar 12 '19 at 20:24