CGAL users discussion list

[Philipp Moeller <>]

Alvin Smith 
<>
 writes:

> Monsieur Rineau,
>
> Thank you for taking the time to respond to my post.  May I point out,
> however, that your response is not very helpful?  I have seen many
> posts on this list that when someone has a difficult question, the
> automatic response is, "Read the full manual," or, "This is beyond the
> scope of CGAL."
>
> No one could agree that having everyone becoming an expert in QT
> drawing routines and writing from scratch methods for drawing CGAL
> objects is a good use of time.  "Well tough bananas!" someone might
> respond, "Why should we support idiots who can't bother to fully
> understand QT?"

Unfortunately this is exactly the case here. CGAL focuses on
Computational Geometry not on visualizations and the demos that come
with it are usually just to demonstrate what you can do with the
library. They are also often more ad-hoc constructs than designed to be
understandable/extensible and that's why their code isn't documented
most of the time. It would be even worse for users to give them the
illusion of having a stable foundation to build on while everything
breaks with every CGAL release.

> There is some truth in this, but there is also a demo (Polyhedron)
> packaged with CGAL that actually does all of this!  My point was that
> this demo is fairly complicated, and not very useful as a tutorial.
> As you see, my question was actually to ask if there was a simpler
> demo, or some other method to to visualize.  A better answer than the
> one you provided would have been, "What you're looking for does not
> exist," or, "Go use meshlab," or even better, "Here's a demo that some
> guy put together awhile ago show someone exactly how to do this."

In general, I agree with you. There should be a more thorough foundation
for visualization with CGAL. Especially given that the compatibility
with visualization tools is rather partial. 

But as you have figured out by looking at the Polyhedron demo this is
far from easy given the widely different requirements between
demos. There is already some effort to conflate the more similar demos
into a single one, but that takes time.

Bottom line: No one is going to invest time into simple drawing routines
because they just don't provide enough on the worth/invest scale. On the
other hand something full-featured is a rather large undertaking and no
one came around to doing it either. 

Well, in the second case you will certainly again end-up with something
that requires deeper understanding and a larger manual.

PS: Please don't top-post.

>> From: 
>> 
>> To: 
>> 
>> Date: Mon, 22 Oct 2012 11:38:13 +0200
>> Subject: Re: [cgal-discuss] Tutorial for visualisation in CGAL using QT
>> 
>> Le dimanche 21 octobre 2012 21:05:26 Alvin Smith a écrit :
>> > Hello everybody,
>> > 
>> > I am trying to visualize a few things in CGAL.  I am definitely a 
>> > beginner
>> > to CGAL, but I've been able to turn a point cloud into a surface using 
>> > the
>> > surface reconstruction routines. At least, I think I have, but I can't
>> > visualize the result, nor can I visualize the point cloud for that matter
>> > without resorting to using Matlab.
>> > 
>> > I'd like to do this with QT, as geomview looks crufty and old, and maybe 
>> > I
>> > have to show my result to a non-technical person and convince them that I
>> > am a serious person.
>> > 
>> > The documentation is insufficient:
>> > 
>> > http://www.cgal.org/Manual/latest/doc_html/cgal_manual/GraphicsView/Chapter_
>> > main.html
>> 
>> That documentation requires that you have read the documentation of the Qt 
>> Graphics View Framework:
>> 
>>   http://doc.qt.digia.com/latest/graphicsview.html
>> 
>> and you should also know about Qt. The CGAL manual does not even try to be 
>> a 
>> tutorial for making Qt demos. You should learn Qt using the Qt manual and 
>> Qt 
>> tutorials. For examples, Qt Graphics View examples:
>> 
>>   http://doc.qt.digia.com/latest/examples-graphicsview.html