Is a tangent to a curve in a hyperbolic plane straight?

Consider a projective plane with an absolute quadric, so that it is a hyperbolic plane.
Given a curve I wonder how the tangent to a curve is defined in a plane with constant positive curvature.

I am unsure whether the tangent in a space of constant positive curvature should be defined as a straight line or somehow as a geodesic and how would I describe such a geodesic geometrically?

When possible, I understand synthetic explanations much better.

• Perhaps this will be something you can use: en.wikipedia.org/wiki/Beltrami%E2%80%93Klein_model – Alan Apr 29 '14 at 19:34
• @Alan In the Beltrami-Klein model, do you think the line connection two points has also the smallest possible (hyperbolic) length, i.e. is it a geodesic line? – Gerard May 1 '14 at 20:56
• Yes, I think the straight lines in this model of the hyperbolic plane are shortest lines in terms of the hyperbolic metric. The model that I've been working with recently gives a construction of geodesics on the hyperboloid which correspond to orthogonal circles in the disk model , – Alan May 1 '14 at 21:45
• I'll leave a drawing of the model, (it's well known to many I'm sure). I was interested in it because it enabled me to produce geodesics on the hyperboloid through two given points. – Alan May 1 '14 at 21:56

One usually defines tangent line $L=T_p(C)$ to a curve $C$ in a surface $S$ at a point $p$ as a 1-dimensional linear subspace of $T_p(S)$. This definition is intrinsic to the topology of the surface $S$. You can also (but this is nonstandard) define tangent to $C$ at $p$ as the unique maximal geodesic in $S$ tangent to the line $L$. This of course requires a Riemannian metric $g$. Constant curvature is irrelevant here as well as the specific realization of $(S,g)$.
• A projective plane P is not a surface is it? What would be the definition of $T_p(P)$? Is the "Riemannian" of a projective plane with absolute quadric known? – Gerard May 2 '14 at 7:46