[x3d-public] [x3d] Goals for V4

[Joe D Williams]

----- Original Message ----- 

Subject: Re: [x3d] Goals for V4
Len and Roy


"As you point out, the event model is one area. "

"out loud I wonder if the current event model in 19775-1 could be
described as ‘synchronous’. While what we need for HTML5 is an
‘asynchronous’ model."

Hi Kids,

The major ideas about the X3D event system is that there exists the
event cascade and the non-looping rule.

Players that observe the cascade will not allow the new frame to be
actually rendered until the cascade, or set of linked cascades, is
complete. This is done by time stamp.

DOM do not due adequate timestamp stuff now, but maybe someday, so for 
now the X3DOM author will need to rig the cascade hisself if it is 
important. This has always been true and is why VRML and X3D have
timestamp and the concept of the event cascade.

This is very important when you are simulating and wish to observe
exactly which frame an event cascade was completed. Players that do
now observe the cascade rule may leak some series of events between
frames.

So, I think ‘synchronous’ must be thought of with respect to completed
frame, not absolute time. An 'asyncronous' system could leak events
between frames making animation appear wrong for a frame or two if the
scenegraph update was ot actually complete. In 'realtime' the next
frame is produced as soon as it is available and the scenegraph is
updated, not on a fixed schedule.

The next major issue is the evaluation sequence that happens after the
current scene is complete and we start preparing for the next. What is
the order of player considerations to produce the next frame. I the 
first one to check if the clock has changed since we last checked?

Thanks and Best,
Joe



<snip>

--------------------------------------------------------------------------------


> _______________________________________________
> x3d mailing list
> x3d at web3d.org
> http://web3d.org/mailman/listinfo/x3d_web3d.org
>