I think I've gone wrong with my reasoning somewere here but I'm not sure why.

We embed $\mathbb{R}$ into the projective plane by $(x,y)\to[1,x,y]$, and consider the projective lines corresponding to $y=mx$ and $y=mx+c, c\neq 0$ and find there intersection.

Each projective lines must be the projection of some $2$ dimensional subspace, so by letting x=0 and x=1 get the lines being the projections of

$span((1,0,0),(1,1,m))$ and $span((1,1,m+c),(1,0,c))$ or all vectors of the form

$(p,q,mq)$ and $(r,s,ms+cr)$

These subspaces intersect when $p=r, q=s$ so that $r=0$ and so the points of intersection are of the form

$(0,s,ms)$ which corresponds to the point in the projective plane $[0,1,m]$.

Is this correct? I'm worried that my answer is independent of $c$.


Yes, you are correct!

The line defined by $y = mx$ in $\mathbb{R}^2$ embeds into the projective line defined by $y = mx$ in homogenous coordinates.

The line defined by $y = mx + c$ embeds into the projective line defined by $y = m x + c z$ in homogenous coordinates.

For $c\neq 0$, these two lines intersect when $y - mx = y-mx - cz = 0,$ that is $z = 0$ and $y = mx$, which describes the point $P = [0:1:m]$.

So, why is it independent of $c$? Well, the two lines you started with in the real spline are parallel to each other. Hence, their intersection in $\mathbb{P}^2$ will be in the line at infinity (i.e., the first coordinate of $P$ is $0$). But, the points in the line at infinity in $\mathbb{P}^2$ corresponds to different slopes of the real lines in $\mathbb{R}^2$. That is, parallel lines pass through the same point at infinity, and the constant $c$ does not affect which point that is.

  • $\begingroup$ Thanks, that's a good explanation. $\endgroup$ – Tim Dec 13 '14 at 0:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.