MinSphereOfSpheresTraits

Definition

A model of concept MinSphereOfSpheresTraits must provide the following constants, types, predicates and operations.

Has Models

CGAL::Min_sphere_of_spheres_d_traits_2<K,FT,UseSqrt,Algorithm>
CGAL::Min_sphere_of_spheres_d_traits_3<K,FT,UseSqrt,Algorithm>
CGAL::Min_sphere_of_spheres_d_traits_d<K,FT,Dim,UseSqrt,Algorithm>

Constants

MinSphereOfSpheresTraits::D
specifies the dimension of the spheres you want to compute the minsphere of.

Types

MinSphereOfSpheresTraits::Sphere
is a typedef to to some class representing a sphere. (The package will compute the minsphere of spheres of type Sphere.) The type Sphere must provide a copy constructor.


MinSphereOfSpheresTraits::FT
is a (exact or inexact) field number type.
Requirement: Currently, FT must either be double or float, or an exact field number type. (An exact number type is one which evaluates arithmetic expressions involving the four basic operations and comparisions with infinite precision, that is, like in  R .)


MinSphereOfSpheresTraits::Cartesian_const_iterator
non-mutable model of the STL concept ForwardIterator with value type FT. Used to access the center coordinates of a sphere.


MinSphereOfSpheresTraits::Use_square_roots
must typedef to either CGAL::Tag_true or CGAL::Tag_false. The algorithm uses (depending on the type MinSphereOfSpheresTraits::Algorithm) floating-point arithmetic internally for some intermediate computations. The type Use_square_roots affects how these calculations are done: When Use_square_roots is Tag_true, the algorithm computing the minsphere will perform square-root operations on doubles and floats where appropriate. On the other hand, if Use_square_roots is CGAL::Tag_false, the algorithm will work without doing square-roots.
Note: On some platforms the algorithm is much faster when square-roots are disabled (due to lacking hardware support).


MinSphereOfSpheresTraits::Algorithm
selects the method to compute the minsphere with. It must typedef to either CGAL::Default_algorithm, CGAL::LP_algorithm or CGAL::Farthest_first_heuristic. The recommended choice is the first, which is a synonym to the one of the other two methods which we consider ``the best in practice.'' In case of CGAL::LP_algorithm, the minsphere will be computed using the LP-algorithm [MSW92], which in our implementation has maximal expected running time O(2d n) (in the number of operations on the number type FT). In case of CGAL::Farthest_first_heuristic, a simple heuristic will be used instead which seems to work fine in practice, but comes without a guarantee on the running time. For an inexact number type FT we strongly recommend CGAL::Default_algorithm, or, if you want, CGAL::Farthest_first_heuristic, since these handle most degeneracies in a satisfying manner. Notice that this compile-time flag is taken as a hint only. Should one of the methods not be available anymore in a future release, then the default algorithm will be chosen.

Access Functions

FT traits.radius ( Sphere s) returns the radius of sphere s.
Postcondition: The returned number is greater or equal to 0.

Cartesian_const_iterator traits.center_cartesian_begin ( Sphere s)
returns an iterator referring to the first of the D Cartesian coordinates of the center of s.