CGAL users discussion list

[Taus Møller <>]

Maybe a section in the Nef_polyhedron_3 documentation on what kernels
offer what would be a good idea. Currently there is a little bit on
the subject, but it is very vague and apparently causes a lot of
confusion. The first mention of kernels in the documentation for Nef
polyhedra is in relation to the infimaximal box and here it just
mentions that an extended kernel is needed to represent unbounded
polyhedra.

After this the only mention of kernels is in the example section. Here
it simply states that Cartesian kernels need a FieldNumberType and
homogeneous kernal need a RingNumberType. If I then click on
FieldNumberType to find out what it is, it lists a bunch of models
including several floating point types along with "C++ built-in number
types ". This leads to the conclusion that i can just use something
like Cartesian<double>. Only very late is it mentioned that the
Exact_predicates_exact_constructions_kernel is best suited for use
with floating point numbers and only in relation to file I/O from OFF
files. It is not explicitly stated anywhere that all the other kernels
only work for Nef polyhedra when used with integer types. This is
extra confusing as Cartesian<double> is otherwise a perfectly valid
kernel and it only breaks down when used with Nef polyhedra

I realize that this might not cause confusion for experienced users of
CGAL who are used to working with the kernels and number types of
CGAL. However, as a new user that picked specifically to use it's 3D
boolean capabilities this caused quite a bit of problems. Judging from
the other posts in this thread and elsewhere, I am not the only one
who have run into this issue.

On 7 October 2015 at 16:58, Sebastien Loriot (GeometryFactory)
<>
 wrote:
> Nef_3 requires a kernel with exact constructions. So usually the best
> choice is Exact_predicates_exact_constructions_kernel.
>
> About the interoperability with Polyhedron_3, the traits class of the
> polyhedron and of the nef should be the same.
> If you have a polyhedron using another traits, then you can convert it
> using for example CGAL/Polyhedron_copy_3.h as follow:
>
> Poly2 target;
> CGAL::Polyhedron_copy_3<Poly1, Poly1::HalfedgeDS> modifier(source);
> target.delegate(modifier);
>
> Sebastien.
>
>
>
> On 10/07/2015 02:31 PM, Merlin Etzold wrote:
>>
>> Dear All,
>>
>> I am working on the inverted problem of Thiago (I am interested what is
>> left after to polyhedrons are subtracted form each other)  and have the 
>> same
>> issues as Thiago and Taus. Some clarification would be really great.
>>
>> Please keep posting about any insights you may get, I will do the same as
>> I go along.
>>
>> Best wishes,
>> Merlin :)
>>
>> -----Original Message-----
>> From: 
>> 
>>  
>> [mailto:]
>> On Behalf Of Taus Møller
>> Sent: 07 October 2015 09:56
>> To: 
>> 
>> Subject: Re: [cgal-discuss] Nef_polyhedron and
>> Exact_predicates_inexact_constructions_kernel
>>
>> I have run into the same problem. I have a hard time know what kernel are
>> usable for what. Even though the exact and inexact construction kernel
>> models the same concepts, they seem to be different in some ways. I have
>> found that the Exact_predicates_exact_construction_kernel
>> works for use with with Nef_polyhedra_3.
>>
>> On 2 October 2015 at 18:30, Thiago Milanetto Schlittler
>> <>
>>  wrote:
>>>
>>> Hello!
>>>
>>>     I’m trying to use Nef polyhedrons to find the intersections between
>>> polyhedrons, and I’m having some problems. The compiler returns the
>>> following error when trying to convert a polyhedron to a nef_polyhedron:
>>>
>>> In file included from
>>> /opt/local/include/CGAL/Nef_3/SNC_k3_tree_traits.h:24:
>>> /opt/local/include/CGAL/Nef_3/Bounding_box_3.h:90:34: error: no member
>>> named 'lo' in 'CGAL::Box_intersection_d::Box_d<double, 3,
>>> CGAL::Box_intersection_d::ID_EXPLICIT>'
>>>        std::copy( q, q + 3, Base::lo );
>>>
>>>     I’m using the “Exact_predicates_inexact_constructions_kernel”,
>>> could this be the reason?
>>>
>>> Thanks in advance,
>>> Thiago Milanetto Schlittler
>>
>>
>>
>>
>> --
>> Regards
>> Taus Møller
>> Software Developer, Apiosoft
>>
>> --
>> You are currently subscribed to cgal-discuss.
>> To unsubscribe or access the archives, go to
>> https://sympa.inria.fr/sympa/info/cgal-discuss
>>
>>
>>
>>
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://sympa.inria.fr/sympa/info/cgal-discuss
>
>



-- 
Regards
Taus Møller
Software Developer, Apiosoft