CGAL users discussion list

[Sylvain Pion <>]

Tom Kazimiers wrote:
> Sylvain Pion schrieb:
> > Tom Kazimiers wrote:
> > > Sylvain Pion schrieb:
> > > > Tom Kazimiers wrote:
> > > > > Sylvain Pion schrieb:
> > > > > > Tom Kazimiers wrote:
> > > > > > > I have an interface which represents access to delaunay
> > > > > > > triangulation.
> > > > > > > Part of this interface should be an iterator which allows me to go
> > > > > > > through all the vertices of the triangulation. As an
> > > > > > > implementation of
> > > > > > > this interface I built an adapter to CGAL which works pretty well -
> > > > > > > except for the iterator.
> > > > > > > From CGAL I can get an Finite_vertices_iterator but I need a
> > > > > > > std::iterator (or a boost iterator) (part of the interface)
> > > > > > > - now I wonder if there is any other way for this then writing a
> > > > > > > wrapper
> > > > > > > class?
> > > > > > It would be interesting to know the reason why you have such a
> > > > > > requirement.
> > > > > I am working currently on an university project in which I would
> > > > > like to
> > > > > easyly switch between delaunay triangulation variants (like between
> > > > > CGAL
> > > > > and an implementation of one of my professors). For this I wrote an
> > > > > interface (lots of "virtual ... = 0;" functions) to write code only
> > > > > against this interface (which is good habbit anyway I guess). This
> > > > > again
> > > > > would it make easy to see the behaviour of different triangulations
> > > > > without changing much code.
> > > > Then I guess the answer is that you have to write wrappers.
> > > >
> > > > You can also go on the generic programming path using a concept,
> > > > and write adaptors for the 2 implementations.
> > > Sylvain, thanks for your suggestions. CGAL makes a lot use of the
> > > concept/template approach, doesn't it?
> > Yes.  You can even call it abuse, sometimes :)
> >
> > You will find the basic principles in the design of the STL.
> Hehe, I see. Sorry, but I got one more littel question regarding the the
> concept thing:
> Concepts seem to be something like a not-written-out interface contract
> - is this true? Like one assumes serveral methods in a instantiated
> template.

This is getting off-topic for this list.
I'd suggest that you take a look at some good book,
like Matt Austern's "Generic Programming and the STL".

--
Sylvain Pion
INRIA Sophia-Antipolis
Geometrica Project-Team
CGAL, http://cgal.org/

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature