metaphorz
YaBB Newbies
http://vam.anest.ufl .edu/wip.html
Posts: 1
|
Howard:
Some thoughts..btw, I don't see any easy way to insert parts of a message
as quoted, but I probably have not explored this BB software completely.
> One question I would like to explore, is generally how to integrate 2D/3D graphics into
> a database along with equations and everything else. Goal is to provide an easier way
> to build 3D virtual world environments driven by dynamic simulations.
It might be worthwhile to determine the specific purpose of the project - is it to: integrate all knowledge
and information together regardless of model type? That is what I am reading in what you
wrote. I share your interest in having a capability to integrate, especially, dynamic models
and geometry models. We have some papers that deal with some aspects of integrating
ontologies with dynamic models, but I know you also are doing a lot of research here.
> What we did so far with the Rinker Building was
> 1. export some existing 3D Cad drawings from 3DMax to VRML
> 2. import VRML files into Lyra objects
> 3. used Xj3D as a platform to render the 3D scene directly from database objects using the Xj3D API > (no intermediate files required)
> 4. we were able to link the 3D shapes to other database info (like building materials), and were about > to incorporate dynamic behavior...
What was the goal of the Rinker Building project? We've used technologies like VRML, Blender,
and Second Life extensively but they have to be taken in light of precisely how they are being
used, and what they bring to the table in terms of capabilities that do not need to be implemented.
> Then SecondLife took off
> 1. unable to import graphics dynanically from a database, or connect to any external
> database from SecondLife
Not being a 'Linden', I am unsure of some of their early design decisions, and some of them
do seem arbitrary or flawed. However, consider that Second Life is built with the idea that all
content is user-generated. This leads to certain design considerations: content must be streamed
in real time from the servers, efficiently and parametrically. So forget arbitrary meshes for the
simple reason that the bandwidth is not sufficient. On the other hand, parametrically defined
surfaces should be possible, and Linden Labs seems not to have explored this avenue (i.e.,
B Spline surfaces). I am not sure why.
> 2. The SecondLife scripting language is virtually useless for doing simulation (we gave up on trying to > autogenerate the scripting language from our model database)
Let's clarify - the scripting language (LSL and transitioning to Mono) is reasonable for 1) discrete event
simulation OR 2) continuous simulation performed on your own server with sync messages to the
Second Life servers (via XMLRPC). For truly real time modeling, Second Life is not the platform
of choice unless one can live with the XMLRPC solution and a frame update rate that depends on
net bandwidth.
> 3. after much frustration, decided it is impossible to implement dynamic 3D simulations in SecondLife > without doing tremendous amounts of work....
True. For real-time stuff, might I recommend something like either Processing (for mostly 2D
work, although JOGL works fine) or OGRE (conjunction with OPENGL). But if you want a platform
that does everything (collaborative building, social collaboration, custom continuous-time dynamics,
robust scripting language, great documentation, mature and large user base), I am not sure that
this exists yet.
We use Second Life for cultural simulation, and for that purpose, it is a good platform.
There are no dynamic real-time needs other than the Havoc physics that comes with
it.
> Then Digital Worlds Institute began to explore another package, called OLIVE
> 1. API written in C++, I'm not too excited about going back to work in C++, but would do so if > necessary
> 2. there are no higher-level authoring tools for OLIVE. We might eventually get it to work, but it would > take lots of work
You might also try a free platform (Multiverse) if the multi-user aspect is important. If this is
not important, try OGRE.
> So... what next?
> - the concept behind Croquet looks interesting, a higher-level programming language integrated with a > 3D presentation (seems similar to Xj3D on the surface). What's not to like about Croquet? (I don't
> know it well enough to say yet). Did you develop a Smalltalk alternative? What do you have in
> mind?
I looked at OpenCroquet two years ago, but it seemed like half-baked technology back
then, and a poor user base. Maybe it has changed, and this could be worth exploring. I'd
give it another go, for sure.
> - I'd like to see an environment where the database stores 3D shapes, dynamic behavior (via
> equations), geometric structure, and automatically builds the 3D interface driven by simulation....
> authoring tools are at the database level, no programming required... but I can dream can't I!!
When VRML was the big thing, I used to dream that someone would start building IDEs and
apps for it, but it never really happend. After all, it is primarily a protocol layer for geometry
and simple node scripting.
It seems to me that you are doing the right thing--you are stating what your key concerns
are and then seeing what platforms support these concerns. It sounds like "real-time"
model behavior is important, so this does limit the choices.
-pau
|
Pages: 1
Integrating 2D/3D graphics into a database (Read 3954 times)
