Fergus Murray wrote:
> That'll be why it looks like the page hasn't been updated for about two
> years then...
Wouldn't surprise me....
> This is pretty much the conclusion I had reached; cheers for the
> confirmation. Do VRML Java scripts allow one to do almost all the same
> tasks as an applet (I'm thinking networking)?
I believe most VRML browsers just ask for the HTML Browser's VM and fire
the code to that. Basically that means the same restrictions as any
applet. See the DIS-Java-VRML working group as they do a lot of
networking + scripts.
> working from Lea/Matsuda/Miyashita's 'Java for 3D and VRML Worlds,'
Wow! Where'd you find a copy of that from? Haven't seen that around for
years. I thought that mine (Late Night VRML) was the last book in print
(it is now out of print). Seriously, if you want to do script work, see
if you can locate a copy of the late night book as at least have the
500+ pages is devoted to the topic.
> happy to switch to it. I never came across any warnings that the EAI was
> going to slow things to a crawl if I used it for something like updating
> twenty-three objects in real time... maybe I wasn't reading between the
> lines carefully enough.
Updating one object in real time is too much for the EAI. The basic
problem is that you have events coming from outside of the runtime scope
of the scene graph. These events arrive at any time and the browser must
find an appropriate time to merge them into the currently running event
cascades. The original spec had the problem (or more correctly, the
defacto implementation due to any other spec'd control) with forcing
another render every time an event was recieved from the EAI. Thus if
you wanted to set a Transform rotation and translation, that would cause
2 frames to render rather than accumulating the results into one. Now,
start considering a humanoid where you are animating a walking motion
and you've now got maybe 50-60 updates, each one causing it's own new
frame render. Hmmm... where did all that performance go?.....
EAI II attempts to solve this problem by allowing the external code to
buffer a collection of changes and inject them all in one hit. However,
it still causes a performance drop as browser still has to queue all the
changes and make sure they get taken at the right time and don't
interfere with the current event cascade evaluations.
> Anyway, I'm working through your 'Raw J3D' tutorial now. I think it's
> going to be worth my while learning Java 3D.
Oh, yeah. I really must find the time to finish that one day....
--
Justin Couch http://www.vlc.com.au/~justin/
Freelance Java Consultant http://www.yumetech.com/
Author, Java 3D FAQ Maintainer http://www.j3d.org/
-------------------------------------------------------------------
"Humanism is dead. Animals think, feel; so do machines now.
Neither man nor woman is the measure of all things. Every organism
processes data according to its domain, its environment; you, with
all your brains, would be useless in a mouse's universe..."
- Greg Bear, Slant
-------------------------------------------------------------------
===========================================================================
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".
