Re: [Fresco-devel] new Path API

To nicholas:

Good to hear that you will be back! I'll leave the Path API for you to
implement: I'm actually a bit scared about all your GL code that does
ignore the coding standard (which wasen't around when most of the code
was written) for the biggest part.

Stefan Seefeld <[email protected]> wrote:
> while nurbs are general enough, often there is hardware (or software) 
> support for
> subsets such as quadrics or bezier. It would be exceedingly hard to detect 
> automatically
> when a given nurbs parameter set maps to a bezier, because the mapping is 
> non-trivial.

A Bezier is a B-Spline with n=p+1 control points and the first p+1 knots
equaling 0 and the last p+1 knots equaling 1. That should be easy enough
to test for. We could provide a hint ("bezier"), so we don't
have to redo those tests all the time... NURBS are B-splines defined on
homogenous coordinates. Does that make the mapping non-trivial?

Rational Bezier curves can describe Quadrics as well. I don't really under-
stand how you are defining Quadrics, so I don't know how to test for those,
but there should be a way to test wether a bezier is a quadric...

> So instead, a user of the API (or a drawing editor) may offer bezier curves 
> directly
> (or quadrics such as circle segments or spheres etc.).

I am not arguing wether the application developer should be able to use bezier
curves. The question was: Do we want to support them on the drawingkit level?
If a drawingkit can accelerate quadric-drawing isen't a "quadric" hint along
the lines of "convex" sufficient?

-- 
Gruss,
Tobias

------------------------------------------------------------
Tobias Hunger           The box said: 'Windows 95 or better'
[email protected]                      So I installed Linux.
------------------------------------------------------------

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQE/NkLgv0FZW3NyoqURAr3cAKCAbLqZ0ynkczJGVeZBLJYpchCsNQCfdZx0
y2hxNVEpTIrG1FvvHh8SoIo=
=/Hz+
-----END PGP SIGNATURE-----