CGAL users discussion list

[Bernhard Kornberger <>]

I'd like to get sure that my interpretation of this kernel
is correct:

typedef Filtered_kernel< Simple_cartesian<double> >
        Exact_predicates_inexact_constructions_kernel;

Can you confirm that:

a) Using this kernel, constructions are done only with
double precision floating point arithmetic
b) Therefore, even exact predicates will possibly fail,
when they use constructed objects like, e.g., in the code
below:

Line l1;
Line l2;
Point p=intersection(l1,l2);
assert(l1.has_on(p));


Thank you,

Best
Bernhard