CGAL users discussion list

[Peter Hachenberger <>]

Hi Andy,

the ply format is better since it is based on the indexed face set idea.
That means that you first store the coordinates of all vertices and then
give the facets as list of references to the vertices. Many file formats
use that---amongst others the off file format, which we use as input. It
is easy to write a converter ply to off (maybe also easy to find on
internet), which would allow you to use our normal interface. 

The stl file format does not use an indexed face set. Instead every face
is a list of vertices, where every vertex is directly given by its
coordinates. The main problem with converting something like that into
an indexed face set is that you want to the coordinates of every vertex
listed only once in the indexed face set. So, that's a bit more
complicated and it gets worse if the data has floating-point problems.

Peter

On Fri, 2008-04-11 at 19:04 +0200, Andy Hirsh wrote:
> Hi Peter
> 
> I sent last mail while logged on my colleague gmail's account... sorry!
> 
> Looking for the rabbit object you talk about in your mail, I found
> that there is another simple 3d format named PLY
> http://en.wikipedia.org/wiki/PLY_(file_format)... is this better for
> the intersection purpose ?
> 
>   best regards
> 
>              Andy
> 
> 
> On 4/11/08, Giampaolo Bellini 
> <>
>  wrote:
> > Hi Peter
> >
> >  thanks for for your quck reply.
> >
> > > with 3D Nef polyhedra it is not a problem to do intersection of planes
> > > and polyhedra. The problem in your case is the file format. We don't
> > > support stl files at the moment.
> >
> > this is not a problem as writing a parser for stl is really a simple
> > task... I have a simple library based on FLEX... could it helps ?
> >
> > > I can write a constructor for stl files
> > > for you in an hour, but it will not be very efficient. It will construct
> > > every facet as a separate Nef polyhedron
> >
> > really ? sound great!
> > I'm just looking at NEF polyhedron as they are totally new for me.
> >
> > > So now the question is, how
> > > big your input data is, and how important the construction efficiency is
> > > for you.
> >
> > actually efficenty is not a problem as I'm still experimenting with
> > STL... and I don't know exactly where I'm gone :-)
> >
> > > Peter
> >
> >  again, thanks a lot for your help, Peter...
> >
> >               Andy
> > --
> > You are currently subscribed to cgal-discuss.
> > To unsubscribe or access the archives, go to
> > https://lists-sop.inria.fr/wws/info/cgal-discuss
> >