Re: Scheper's Catmull-Rom curves, and Spiro curves

On Wed, 2010-07-14 at 04:08 +0400, Alexandre Prokoudine wrote:

> The problem with existing implementation of Spiro in Inkscape is that
> it is implemented as a live path effect, and LPEs have a significant
> disadvantage: you never see what you created until you stop drawing.
> This badly affects both Spiro Spline and Pattern Along Path effects
> which, being artistic effects, are supposed to be interactive.
> 
> Another real problem with Spiro as LPE is that you can't apply boolean
> operations. There is a badly working LPE for boolean operations that
> is currently not built by default, does only few simple operations,
> crashing Inkscape half the time anyway and is conceptually clumsy.

The curve between two adjacent nodes on a Spiro curve by necessity
depends on other nodes in the path. This makes it difficult to implement
Boolean operations on Spiro curves. For example, in cutting a path into
two pieces you are both adding (the new end nodes) and removing nodes
from each path. This is going to be true of any curved path that is
defined only by nodes on the path. You can try cutting Spiro paths in
two with FontForge and see how the shapes change.

     Tav

Received on Wednesday, 14 July 2010 09:32:57 UTC