Skip to Content.
Sympa Menu

cgal-discuss - Re: Re: Re: [cgal-discuss][Nef_polyhedron]A more complex butstillsimplequestion

Subject: CGAL users discussion list

List archive

Re: Re: Re: [cgal-discuss][Nef_polyhedron]A more complex butstillsimplequestion


Chronological Thread 
  • From: Peter Hachenberger <>
  • To:
  • Subject: Re: Re: Re: [cgal-discuss][Nef_polyhedron]A more complex butstillsimplequestion
  • Date: Wed, 06 Feb 2008 13:07:41 +0100

I might have asked you this before, but why do you need halfspaces? If
you only need halfspaces to perform intersections, then you don't need
the extended kernels. I wrote an intersection routine that intersects a
halfspace with a Nef polyhedron without constructing the halfspace
explicitly.

Peter

On Wed, 2008-02-06 at 09:01 +0800, Max wrote:
> Hello Peter
>
> From what I get from your mail msg's, AFAIUC, there's
> the following facts with relate to Nef_3:
>
> 1. To use half-space nef_3 ctor such as
> Nef_polyhedron_3 N(Plane_3(1, 0, 0, -1), Nef_polyhedron_3::INCLUDED);
> one should use extended kernel;
>
> 2. Don't use inexact kernels with Nef_3;
>
> 3. To play with Nef_3 with double type coordinates, one should
> use SNC_indexed_items as the 2nd template parameter in instantiating
> a Nef_3 type;
>
> 4. Extended kernels do not work together with SNC_indexed_items.
>
> It seems that there's no choice, of combinition of Nef_3/kernel, for me
> to meet these needs:
>
> a) I need the half-space Nef_3 ctor; (that means I should used extended
> kernel)
>
> b) my coordinates is double type; (that means I should use
> SNC_indexed_items, but it's
> incompatible with extended kernel)
>
> So this seems to be a deadlock. Is there any solution?
>
> Many thanks for your help, Peter.
>
> B/Rgds
> Max



Archive powered by MHonArc 2.6.16.

Top of Page