I've just started to learn about Java, but... The approach you described for
these supplemental rendering engines sounds awfully similar to the whole Java
notion of downloading "applets" on the fly. At some point I assume that if a
certain type of applet becomes sufficiently prevalent it will be incorporated
into future versions of the browsers.
I'm trying to develop a mental model of how Java and VRML will fit together in
the (hopefully) near future. The notion of using the VRML scripting language
to describe these brwoser extensions is also a good one. But, as an open
question to the community, I'd like to find out if anybody knows the state of
scripting/animation supplement to VRML and the nature of that language.
Jim
--------------------------------------
Date: 10/5/95 12:24 PM
To: James Rayson
From: Tom R Craver
Received: by qmgateib.mitre.org with SMTP;5 Oct 1995 12:21:02 -0500
Received: (from root@localhost) by mbunix.mitre.org (8.6.10/8.6.9) id MAA11979
for [email protected]; Thu, 5 Oct 1995 12:18:49 -0400
Received: from tres.wired.net (tres.wired.net [204.62.130.122]) by
mbunix.mitre.org (8.6.10/8.6.9) with ESMTP id MAA11964 for
<[email protected]>; Thu, 5 Oct 1995 12:18:47 -0400
Received: from get.wired.com by tres.wired.net (8.6.12/8.6.5) with ESMTP id
JAA22066; Thu, 5 Oct 1995 09:28:59 -0700
Received: by get.wired.com (8.6.12/sven) id JAA11322; Thu, 5 Oct 1995 09:15:51
-0700
Received: by get.wired.com (8.6.12/sven) id JAA11271; Thu, 5 Oct 1995 09:15:43
-0700
Received: from ormail.intel.com by get.wired.com (8.6.12/sven) with SMTP id
JAA11266; Thu, 5 Oct 1995 09:15:42 -0700
Received: from relay.jf.intel.com by ormail.intel.com with smtp
(Smail3.1.28.1 #7) id m0t0stZ-000UmBC; Thu, 5 Oct 95 09:11 PDT
Received: from ccm.hf.intel.com by relay.jf.intel.com
(Smail3.1.28.1 #2) id m0t0stY-000twnC; Thu, 5 Oct 95 09:11 PDT
Received: by ccm.hf.intel.com (ccmgate 3.2 #3) Thu, 05 Oct 95 09:11:40 PDT
Date: Thu, 05 Oct 95 08:57:00 PDT
From: Tom R Craver <[email protected]>
Message-ID: <Thu, 05 Oct 95 09:11:00 [email protected]>
To: [email protected]
cc: [email protected]
Subject: CSG+D, generalized cylinders, NURBS, extrusion, etc
Sender: [email protected]
Precedence: bulk
X-MDF: Mail for jrayson sent to [email protected]
Perhaps VRML should allow a 'plug-in parser' approach. Allow application
specific extensions to the VRML modeling language could be created, with
their own scripts embedded in a special section. This would dispense with
the need for any 'smart modeling' extensions like CSG+D, generalized
cylinders, extrusion, revolution surfaces, NURBS, quadratics, etc.
If a browser lacks the code to parse a section, it would be represented by a
stand-in 'bounding box' and label, which the author (or generating tool) is
required to specify. It should be possible to distribute the parser
extension via URL esference in the VRML file, installing it just before it
is needed. The user might be given the option to install it for future use
- otherwise it'd be discarded on exit.
One way to try out new parsers would be to use the same machine independent
scripting language used for behaviors. Some such extensions will become
popular and be offered in fast machine-specific form.
It should be possible to assign standard parser types to such extensions,
and allow authors to specify such a type as well as a 'fall-back' URL in
case the user doesn't have the correct machine dependent parser installed.
If a type hasn't been standardized and 'installed' in the browser, its scope
would be limited to the scope of the world it is defined in.
Someone is likely to suggest using a URL to esfer to a script file, instsad
of a special embedded section. That might work, so long as a single
'extension script' file type is used, with the parsing-helper specified in
the main VRML file (so the parser can avoid downloading the script file if
it lacks the correct parser for some reason). But often the point of these
extensions is to be very compact - small enough to fit well in-line. I tend
to favor allowing in-line 'special parser' script, and letting the parser
put a URL in that section if it wishes.
Tom Craver
Intel Corporation
Opinions Expressed are Mine, Not Intel's