X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f X-Recipient: geda-user AT delorie DOT com Subject: Re: [geda-user] Router fun with rubberbands From: Stefan Salewski To: geda-user AT delorie DOT com In-Reply-To: References: <1354748372 DOT 3386 DOT 3 DOT camel AT AMD64X2> Content-Type: text/plain; charset="us-ascii" Date: Fri, 07 Dec 2012 18:10:05 +0100 Message-ID: <1354900205.3512.34.camel@AMD64X2> Mime-Version: 1.0 X-Mailer: Evolution 2.32.3 Content-Transfer-Encoding: 7bit Reply-To: geda-user AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: geda-user AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk On Fri, 2012-12-07 at 14:34 +0100, Kai-Martin Knaak wrote: > Stefan Salewski wrote: > > > I just fixed some trivial bugs of my toy router and put some pictures on > > my page: > > > http://www.ssalewski.de/Router.html.en > > Cool. > Just curious: Is there any reltion to the topo router project that was > started during Google summer of code? > > I hope, your project is going fly! > > ---<)kaimartin(>--- There are only some indirect relations to Anthony's router: -- I was impressed by his pictures -- I use GTS library for delaunay triangulation as he did Thats all -- I looked some hours at his code, but I was only able to understand very small portions. The main problem for me was, that I was not able to see relations between his code and the papers. He made much use of functions of the GTS library, while I use that library only for the initial delaunay triangulation. Indeed, it was my impression that he followed not very close the papers and the PhD thesis of Tal Dayan. I think that Anthony is a very smart person and he has coded much of his own ideas. One problem is that low level C language makes it really very difficult to develop and debug such type of router. Currently I am investigating switching from GTS library to CGAL. GTS is plain C, so it was easy for me to code a Ruby interface for the Delaunay part. But GTS mailing list is inactive since May 2012 -- my subscription failed, and the original authors is not really interested. GTS is ok for me, I do only need the initial constrained delaunay triangulation, exactly the neighbor vertices for each vertex. But I think I will have a look at CGAL in the next weeks, its a large C++ library. Bindings to other languages are not really well supported, there is some basic support for Java and Python. I read that CGAL makes use of templates, which makes it difficult to build bindings which support all the functionality. I don't know enough about C++ to build direct bindings (Java and Python uses SWIG for the bindings), but with some plain C glue code it should be possible. I hope there is nothing wrong with the CGAL license, seems to be dual GPL and commercial. That should be OK, but there is a strange fork project site available: https://gforge.inria.fr/projects/cgal/ Maybe it is only a mirror with GPL license, I have to investigate that. Best regards Stefan Salewski