[GSoC]Idea

[Maurizio Paolini]

On Mon, Apr 08, 2013 at 06:39:42PM +0200, miniBill wrote:
> 2013/4/8 Maurizio Paolini <paolini at dmf.unicatt.it>
> 
> > Going from 2 to 3 coordinates is a big change, we need
> > a broad agreement before proceeding.
> >
> Most definatively.
> On the long term I think the best thing would be having
> all algorithms that can work on generic spaces do so.

Actually, a 3D version of kig is in my opinion a completely
different beast, which should be implemented as a completely
different project (kig3d?) and starting more or less from
scratch.  I wouldn't vote in favour of a modification of
kig with builtin 3D capabilities, actually.

On the contrary, implementing projective coordinates (with
3 homogeneous coordinates in place of the present 2) seems
feasible.

On noneuclidean (2D) geometry I do not presently have a precise
opinion, but I am a little bit uneasy on the matter.
Of course it is already possible to implement hyperbolic
geometry (e.g. Poincare' disk or Poincare' half-plane) as a
model within euclidean geometry, implementing macros like
"hyperbolic line", "hyperbolic segment" and the like.
This works actually surprisingly well and I could provide details
on it.

Maurizio