Hello Praveen,

> Thanks for your prompt reply and useful answers.
>
> > we are still developing software with OpenGL Optimizer as well as
> > OpenSG. Typically you have to rewrite your 3D Layer completely
> > since not all concepts are the same, but its not that hard and
> > OpenSG offers
> > some things not possible in Optimizer.
>
> Hmm... That's an interesting thought but we want to completely move
> away from Optimizer going forward. We are exploring whether to use
> OpenSceneGraph or OpenSG for our application. Since our application
> is CAD related, OpenSG seems to be ahead in the race so far :-)

You might want to take my answer with a grain of salt, I am slightly 
biased towards OpenSG. In general we like OpenSG being more industry 
focused with things like its compact and backward compatible file 
format OSB, the rendering features / speed, the incredible cluster 
support, etc (we are using it for approx 5 years now and never regret 
that decision). I have to admit that I have never used OpenSceneGraph 
so I am not going to comment on it.

> > Not available today. Maybe Dirk can say more about it, its more of
> > a conceptual decision (OpenSG is much more a rendering framework,
> > and sensors and routing belong more into application territory).
>
> OK. I guess, this should be something we can implement at the
> application level then.
>
> > OpenSG has support for NURBS but it is not comparable with the
> > tesselators of Optimizer. The reason is that there is no real
> > support for topologic reconstruction (instead some rendering based
> > concept called fat borders has been used) which makes it completely
> > unusable for us and many other users I have talked to in the past,
> > simply because geometry is not only used for rendering, but also
> > for collision
> > detection and other purposes wherenon-closed gaps can not be
> > tolerated.
>
> Hmm... This part seems more complex to handle. I wonder how much work
> it would be to replicate similar functionality in OpenSG.

Well our hope was that this functionality would have been within the 
NURBS framework and there was some kind of topologic reconstruction in 
early works (called sewing) but due to unknown reasons the implementors 
choose to continue with a rendering based approach. But since it's free 
I can't complain, can I ?

In general writing a topologic reconstruction is not that easy and 
regarding tesselation strategies there more than the one used in 
Optimizer.

If OpenSG would have a tesselator comparable to the one of 
Optimizer/Performer ... well ... killer toolkit.

> > The simplifiers in Optimizer simply suck, because they are totally
> > outdated and the parameters do not allow real control over
> > the quality
> > of the resulting geometry.
>
> I assume you mean that OpenSG has better simplifiers than Optimizer?

Sorry no, OpenSG has no simplifiers yet, but you are of course invited 
to contribute one after implementation ;-) I just wanted to point out 
that those in Optimizer are in practise of no use.

> > Not available, but it would also suffer from the problems with the
> > tesselation if we are talking about NURBS and stuff.
>
> Thanks a lot!

Your welcome

Matthias
-- 
+---------------------+----------------------------+
| VREC GmbH           |                            |
| Matthias Stiller    |                            |
| Robert-Bosch-Str. 7 | tel:    +49 6151 4921034   |
| 64293 Darmstadt     | web:    http://www.vrec.de |
| Germany             | mail:   [EMAIL PROTECTED]         |
+---------------------+----------------------------+


_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to