Many thanks to Mark for the information. I am trying to write some simple 
triangulation code myself, which is optimized by some curvature.
Hope it works for my simple restrained input polygons.

regards,

---white

On Fri, 14 Feb 2003 15:55:56 -0800, Mark Hood <[EMAIL PROTECTED]> wrote:

>> Date:         Fri, 14 Feb 2003 14:28:21 -0700
>> From: White Morph <[EMAIL PROTECTED]>
>>
>> If Java3D does use FIST code for triangulation, then I guess it doesn't do
>> the sorted or random optimization that is mentioned in the paper, because
>> each time I run a program, it always get the same triangulation, on the other
>> hand, if I change the starting vertex, it normally gets a different
>> triangulation.
>
>GeometryInfo uses a Triangulator constructor that sets the sorted and random
>ear modes to false.  I don't know the reason behind this choice; there may have
>been some problem with those modes that it made it not worth using for Java 3D.
>Paul might have some insight into this.
>
>If you think this might be a bug, the best thing to do is submit a bug report
>so we'll have some justification for looking into this complex bit of code.  Is
>there a problem in your application that would definitely be solved by using
>sorted or random ear optimization?
>
>Triangulator is probably not using the latest version of FIST.  It appears to
>be using code derived from FIST 1.6.  The last update to the Java 3D derived
>code was a bug fix on February 25 2002.
>
>> By the way, I wonder why Java3D doesn't use the Constrained Delaunay
>> Triangulation, which even the author of FIST admits has better triangulation
>> quality.
>
>Good question.  We've been using the FIST code since early 1998.  It was
>probably a matter of what code was already available from the relationships
>that we had at the time vs. the resources that we had to develop a robust
>triangulator ourselves.  Triangulators are notoriously difficult to get right
>for "industrial-strength" applications.  Again, Paul might have some insight
>into this.
>
>The Triangulator class is a utility, not part of the Java 3D core, so if you
>need to use an alternative triangulation algorithm there's no technical reason
>why you can't substitute whatever you need.
>
>-- Mark Hood
>
> ==========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff JAVA3D-INTEREST".  For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to