CGAL users discussion list

[Monique Teillaud <>]

Dennis Endt wrote:
> Very Good joke. You made my day ;)

quite good indeed ;)

> But if you had tried to understand 

I am afraid we have spent more time reading your e-mails, trying to 
understand them, and answering them, that you have spent reading the 
manual, reading our answers and trying to understand them... after some 
threshold, mechanically, answers become less appropriate...

> you would know, that I`m only
> searching for a method which tells me, if the triangulations gone wrong
> because of the same x_y coordinates and so on.

Bernd's suggestion about PCA sounds worth trying.

Of course you are welcome to keep on posting on this list, if you 
encounter real problems that are not solved by a simple look at the manual.

BTW, a complement about my answer below: in fact triangulations will 
never crash, but if you use a traits class that is not adapted to the 
data, your points may look collinear in projection, and you will again 
not like the results you get.

best,

> Andreas Fabri schrieb:
> > Dennis Endt wrote:
> > > Monique Teillaud schrieb:
> > > > If you are lucky (= your points really don't belong to a plane
> > > > parallel to any xy,xz,yz plane), you will get the same triangulation
> > > > three times, which is not what you want (I guess so). Yhen using one
> > > > of these random traits should give you what you want.
> > > > If you are less lucky, one of the computationa will crash. Then, use
> > > > one of the other two.
> > > > I'm afraid that's the best you can do if you don't know anything about
> > > > your data, except the fact that they are almost coplanar.
> > > >
> > > > best,
> > > >     Monique Teillaud
> > > Hi,
> > >
> > > thanks so far for your help.
> > >
> > > My last question is:
> > > How can I check if a computational will crash? What is the best method
> > > for this?
> >
> > You should
> >
> > #include <CGAL/Cristall_ball.h>
> >
> > It provides a constructor with an unsigned int specifying the number
> > of milliseconds you want to look into the future.
> >
> > It's an undocumented feature, but that is what we recommend for
> > avoiding planes to fall from the sky.
> >
> >
> >
> > > I tried to use is_valid(), but this didn`t worked. Is it enough to check
> > > for the number_of_vertices != 0 of a fassade?
> > >
> > > Best,
> > > Dennis