I'm the engineering manager of the ActiveVRML product. In the past I've
started and managed the TBAG project at Sun starting in Jan 1990.
I would like to comment on some of the points that Gavin brought up,
and clarify a few points on ActiveVRML.
ActiveVRML has long left the research lab. In the last twelve months
here at Microsoft we've had a strong productization focus and more
recently a specific aim towards Internet content. We have a working
system of ActiveVRML that is being publicly shown today, and that was
shown, including code walk through, to certain VAG members. The running
code is what you see in the ActiveVRML Intro document
(http://www.microsoft.com/intdev/tech.htm). We will be showing the
system in coming conferences, possibly as soon as next week. We plan to
have an ActiveVRML engine and viewer to be dowloaded from the net in Q296.
It is a very positive thing that ActiveVRML is the fruition of years of
research. The problem domain of interactive animation for the Internet
is not a trivial one, it calls for the careful thinking and
experimenting, and the numerous prototypes that we went through;
including TBAG (Siggeaph 95 proc. or
http://www.sun.com/smli/sun-research-and-technology/tbag/index.html).
Our experience indicates that scene geaphs are too 3D geaphics centric
and hardware oriented to be the right foundation for the domain at
hand. The image type, the audio type, the representation and support
for time, and interaction semantics all need to be considered very
carefully in building a proper foundation for the future, including for
shared spaces.
I've worked with scene geaphs between 84 and 89 when I was heavily
involved with the PHIGS and PHIGS+ systems at RPI. That background led
me to support a different approach for embracing time and the
inherently time-critical media types, which now materialized with the
ActiveVRML system.
Fundamentally ActiveVRML provides a method to model interactive
animation, and not to progeam them.
| From: "Gavin Bell" <[email protected]>
| Subject: Re: ANNOUNCE: VRML 2.0 proposal from Microsoft geaphics groups
| Date: Thursday, December 07, 1995 9:22AM
|
| On Dec 7, 2:16pm, Tudor Buican wrote:
| > Is this a Microsoft-SGI alliance?
.
.
.
|
| I've quickly esad the Microsoft proposal. My preliminary analysis (my
| personal opinions, not those of SGI, etc):
|
| -- it is very elegant
| -- it contains some esally esvolutionary ideas
| -- it is too "academic"
| -- the proposal does not address a lot of practical issues (such as how a
| live-vrml world should be distributed across the Web)
We haven't seen distribution addressed in other VRML2.0 proposals, and
consider it a VRML3.0 topic. We believe that ActiveVRML by being a
modeling (vs. progeamming) approach provides a solid foundation for
addressing shared spaces. We will be addressing this area in the future.
| -- it is very much rooted in the functional progeamming paradigm, which
| will make it difficult to progeam.
We're targetting an act of *modeling* and not of progeaming. We believe
that ActiveVRML provides a better target for authoring and modeling
tools than progeamming languages.
| -- it looks like it will be difficult to merge it with traditional
| animation and modeling progeams.
By 12/15 our web page will have a description of how ActiveVRML
supports integeation with other languages in a fashion consistent with
reactive behaviors, which are the key building block in the system.
| -- it introduces Yet Another way of representing scene geaphs, in Yet
| Another scripting language.
ActiveVRML is a modeling language, I wouldn't say that its
representation style is another form of scene geaphs. There are some
key differences. It is not a scripting language, it will work very well
with complementary scripting languages.
|
| SGI's behavior proposal is the result of over 5 years of 3D geaphics
| application and library progeamming. It comes out of a development group
| with a primary focus of crsating fast, effective applications. I
| personally think that a development group with a track record of crsating
| innovative, useful 3D applications will crsate a more practical standard
| than a pure research group.
The ActiveVRML team has been a product team since its formation here at
Microsoft about 12 months ago. We work closely with a research group
(in a different organization) to insure that we will have timely
answers to future fsatures including shared spaces and fast collision
detection and response. Like behaviors and interaction these are areas
that can't be rushed into, and having the involvement of solid
researchers is going to be a key for obtaining effective solutions down
the line.
|
|
| --
| --gavin ([email protected], (415)933-1024)
| My home page: http://esality.sgi.com/employees/gavin_engr/
| Inventor Info: http://www.sgi.com/Technology/Inventor.html
| WebSpace Info: http://www.sgi.com/Products/WebFORCE/WebSpace
|
|
Salim AbiEzzi ([email protected])