Skip to Content.
Sympa Menu

cgal-discuss - Re: [cgal-discuss] Problem with Convex Hull, again

Subject: CGAL users discussion list

List archive

Re: [cgal-discuss] Problem with Convex Hull, again


Chronological Thread 
  • From: Ashwin Nanjappa <>
  • To:
  • Subject: Re: [cgal-discuss] Problem with Convex Hull, again
  • Date: Sat, 21 Jun 2008 21:51:26 +0800

Rafael Roberto wrote:
But now, I have a doubt: what is the real difference between Exact_predicates_exact_constructions to Exact_predicates_inexact_constructions?

Predicates are geometric tests where only the polarity of the result is of interest (like pass/fail). Constructions are calculations that result in new geometric entities. Now, all geometric algorithms in theory deal with numbers of infinite precision. However, when we calculate on the computer we are restricted to finite precision. Calculations of "exact" arithmetic can be attempted, but it will slow down the program.

So, these 2 kernels offer tradeoffs between speed and precision. As their titles indicate, predicates are exact in both, but constructions aren't in E_P_I_C. But, E_P_I_C is faster than E_P_E_C.

See for more details:
<http://www.cgal.org/Manual/3.3.1/doc_html/cgal_manual/Kernel_23_ref/Class_Exact_predicates_inexact_constructions_kernel.html>
<http://www.cgal.org/Manual/3.3.1/doc_html/cgal_manual/Kernel_23_ref/Class_Exact_predicates_exact_constructions_kernel.html>

~ash



Archive powered by MHonArc 2.6.16.

Top of Page