LAW,
Please have a look at the ISO/IEC-14772-1 section 4.12.1 and tell
us why one has to implement JavaScript to be conformant to the
standard.
http://www.web3d.org/technicalinfo/specifications/vrml97/part1/concepts.html#4.1
2
I'll paraphrase, no scripting language is required by vrml97 and
if you do implement one of the suggested scripting languages it must
conform to the definition of the suggestion found in its respective
annex.
So lets at least try to separate the standard from a sample from
some prevalent content. Under these conditions, you simply can't expect
that content generated from browser b with b's favorite scripting
language must absolutely work on anybody besides b's browser.
The work going into X3D incidently, with specific reference to the SAI
http://www.web3d.org/TaskGroups/x3d/sai/SceneAuthoringInterface.html
attempts to bridge the external scripting of web pages with
the internal scripting of the scene by basing the scene elements in
the document object model (DOM) explicitly, essentially using
bean patterns derived from the x3d schema. The hopes are that
at least spelling it out this way disambiguates the semantics
from the syntax of the script language uniformly.
Regards,
______________________________________
Rick Goldberg
Sun Microsystems
901 San Antonio Road, MS UMPK27-302
Palo Alto, CA 94043-4900
650 786-0108 Direct
650 856-2114 Fax
[EMAIL PROTECTED]
> MIME-Version: 1.0
> Date: Wed, 14 Mar 2001 12:37:53 +0800
> From: #LAW YEE WEI# <[EMAIL PROTECTED]>
> Subject: [JAVA3D] Java 3D-VRML97 Loader - Brief sharing of result
> Comments: cc: "Chan Kai Yun, Tony (Assoc Prof)" <[EMAIL PROTECTED]>
> To: [EMAIL PROTECTED]
>
> Personal experience
> -------------------
> I wish there's a more convenient name for the loader even though it's
> already defunct. I would briefly call it "the loader" in the following. My
> experience with the loader has been very frustrating. First, nodes supported
> are kept on a functional level. Second, the parser has serious limitation.
> Third, there's no JavaScript implementation (well, this is actually
> equivalent to the first point). Fourth, the source code is a mess. Fifth,
> performance is unacceptable. Sixth, navigation and interaction is awkward.
>
> Results in a sketchy form
> -------------------------
> After a year of grappling with the code, I have come out with my
> modification that can parse relatively complex scene, supports JavaScript,
> has better text support etc. As a matter of fact, my intention is to
> implement Living Worlds. However the outcome is not what I expected, as can
> be seen from the screen shots below:
>
> http://members.tripod.com/ywlaw/vr/bee-tilted-wave.png
> http://members.tripod.com/ywlaw/vr/bee-tilted-wave-back.png
>
> The screen shots clearly show a wave plane gone awry -- reason unknown.
> Apart from this visual artifact, the performance is disappointing. I've had
> to disable all animation to get a smooth interaction. Of course it can be
> argued that my modification has deteriorated its initial performance to its
> current level, but since the initial loader isn't able to parse the sceen at
> all, comparison can't be made. (The reason that the initial loader can't
> parse the scene is due to the heavy use of PROTOs, JavaScript in the scene.)
>
>
> Thoughts
> --------
> While there have been a lot of talks on advanced graphics techniques like
> collision detection, particle system on the mailing list, none of these have
> ever made their way to the loader. I fully realize all development has gone
> to X3D, but the VRML97 part of X3D has never seen any improvement as far as
> I can tell. While there have been a lot of questions on the use of the
> loader, there hasn't been any significant effort to try explaining the cold
> hard facts -- that the loader isn't ready for any serious application. Of
> course, tagged as a reference implementation, we shouldn't expect much from
> it. Therefore I'm tempted to call for an alternative open source
> implementation. In fact, since there have been calls for an open source
> implementation of Java 3D itself, why not a VRML loader?
>
> Footnote
> --------
> (1) The scene is a part of the Tpresence Holodesk package
> (http://www.holodesk.com/hd/download/index.html), a virtual world called
> Avalon Reef.
> (2) The navigation interface contributed by Laszlo Seregi
> (http://www.external.hrp.no/vr/) is excellent, thanks Laszlo.
>
>
>
> Best regards,
> LAW, Yee Wei
> http://www.ntu.edu.sg/camtech
> Tel: +65 790 4201
>
> ===========================================================================
> 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".
