CGAL users discussion list

Quoting Michael George 
<>:

> Greetings,
>
> I've been struggling for some time now to extend the minkowski sum code
> in CGAL to produce non-regularized output (i.e. to retain the low
> dimensional features).  I've explored a variety of designs to perform
> the union of cycles, but it seems to me like the arrangement package is
> just slightly too inflexible to accomodate each design.  I have two
> questions:

I guess it is a matter of opinion, but I would say that the  
arrangement package is a good choice to accomplish your task. Besides,  
what do you mean by 'each' above?

> 1. Has anyone managed to do this?  If so, how?

Do what exactly? The Minkowski sum package supports two different  
methods to compute the exact Minkowski sum, one uses decomposition and  
union, and the other is based on convolution. The latter may already  
handle the low dimensional features. The former does not, because it  
uses the union operation, which does not, so the problem reduces to  
computing the non-regularized union. What type of boundary curves are  
we talking about?

> 2. If not, I'm probably going to have to implement my own sweep-line
> algorithm to do the merging, instead of using the arrangement package.
> At that point I would probably break away from CGAL entirely and just
> implement everything from scratch for simplicity.  I'm wondering if
> this would still be considered a derived work for licensing purposes,
> since I wouldn't have any CGAL code but would still be using the design
> of CGAL and the minkowski sum algorithm in it as a guide?  I'd like to
> license my code under the GPL which is incompatible with the QT license.

That depends on how close to the original code your code is going to  
be, and whether your final code will absolutely not have any sign of  
the original code.

The convolution algorithm and the decomposition algorithm to compute  
Minkowski sums where published and appear in the literature.

> Thanks,
>
> --Mike
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss