Hi all,

from all the discussion going on since the realease of 1.3 beta I would
summarize that the transparency problem solves some problems now but
still not all. For me it is therefore not solved. It seems that it will
only be (completely ?) solved if a triangle sorting mechanism exists.
Since users of the 3d API have no access to the inner structures, any
code written by the user will be slower than if it has access to inner
structures and can arrange them for efficient sorting. Furthermore the
sorting has to be done for every Canvas3d showing the
same scene (this is also a reason why e.g. an OrderedGroup will not
solve the problem). This can only be done in mixed mode rendering which
I think will slow everything even further.

I would request an extension of the j3d API with a triangle sorting
capability additionally to the geometry sorting capability. Since the
sorting time depends very much on the number of objects to sort, i think
it will be usefull for cases with a limit number of objects. If it is
combined with geometry sorting it might help the users to split their
geometries maybe more efficient because the splitting requirements are
softened. Furthermore the documentation can state that there will be a
drop on performance (since the machines get faster every year this
problem might vanish in some years).

Any other opinions on that ?

regards
Paul

===========================================================================
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