VRML - Request for Advice: VRML, X3D, Xj3D, Architecture

Author

Request for Advice: VRML, X3D, Xj3D, Architecture

Keith Bennett

2005-05-04, 7:25 pm

Dear Colleagues -

I am new to VRML and 3D programming in general, and would like to ask
you for some advice. We are designing and building a new application
that would display a 3D representation of objects in a data base and
allow various user operations on the objects being displayed (e.g.
query
detailed info, show info about the relationship between two selected
points). We have control over all steps along the way (the DB, the
middleware, and the viewer), so we have a lot of freedom of choice.

Some of these questions are interrelated.

1) Choosing a Viewer

We'd like to avoid a lot of coding, so it would be nice to get a viewer
that has already been written. We'll need something that's customizable
too. It would be nice to get something free, OS-neutral, and open
source
if possible. Xj3D sounds appealing, but it seemed to slow down my
system
tremendously when I ran the sample data, though. Has anyone
successfully
relied on Xj3D for real world use? Do you have any other suggestions?

2) Scripting, and Choosing a Data Format (VRML, X3D, ???)

It would be great to be able to script in Java, Python, or Ruby.
According to http://vrmlworks.crispen.org/eai_faq/scripts.html, VRML
supports two kinds of Java scripting, Java Scripting Authoring
Interface
(JSAI) and External Authoring Interface (EAI). Only the EAI supports
using events and fields that are not part of the script node, so that
is
what we would probably need to use. It says that EAI "is available in
the two major browsers". I don't know when this was written; is it
supported in IE and Firefox? Any other advice?

I also read that scripting had changed significantly from VRML to X3D.
How do the two compare?

3) Browser vs. Application for the Viewer

Any suggestions as to whether to implement the viewer in a web browser
as opposed to a standalone application? I'm thinking a standalone app
would provide greater control and flexibility, and could include Xj3D
or
be a modification of it. Also this would eliminate the problem of
inconsistent browsers. On the other hand, using a browser gives us free
access to some services (such as printing) that would be nontrivial to
implement ourselves. What do you think?

4) Middle Layer - EJB?

We're thinking there would be a piece that would respond to the
viewer's
requests to read the data from the data base (always or almost always
Oracle), and generate output that would be used as input by the viewer
(X3D, VRML, etc.) One possibility would be to use something like JBoss
and use session beans for the interface between this piece and the
viewer. Any suggestions?

* * *

Thanks for taking the time to read this. I appreciate any assistance
you can offer.

- Keith Bennett

jeffs

2005-05-05, 7:27 pm

Keith Bennett wrote:

> I am new to VRML and 3D programming in general, and would like to ask
> you for some advice. We are designing and building a new application
> that would display a 3D representation of objects in a data base and
> allow various user operations on the objects being displayed (e.g.
> query
> detailed info, show info about the relationship between two selected
> points). We have control over all steps along the way (the DB, the
> middleware, and the viewer), so we have a lot of freedom of choice.

sensible use of 3D for visualization
JDBC makes connectivity pretty easy
[see work by "DataBase Working Group" @ web3d site]

> 1) Choosing a Viewer
>
> We'd like to avoid a lot of coding, so it would be nice to get a
viewer
> that has already been written. We'll need something that's
customizable
> too. It would be nice to get something free, OS-neutral, and open
> source if possible.

what is your time-frame?
have you considered xVRML and Carina?
[http://www.xvrml.net/]

> 3) Browser vs. Application for the Viewer
>
> Any suggestions as to whether to implement the viewer in a web
browser
> as opposed to a standalone application? I'm thinking a standalone app
> would provide greater control and flexibility

that is true

> On the other hand, using a browser gives us free
> access to some services (such as printing) that would be nontrivial
to
> implement ourselves. What do you think?

would need to know more about
why printing etc is so difficult for you to implement

> 4) Middle Layer - EJB?
>
> We're thinking there would be a piece that would respond to the
> viewer's requests to read the data from the data base
> (always or almost always Oracle), and generate output
> that would be used as input by the viewer
> (X3D, VRML, etc.) One possibility would be to use something like
JBoss

that might actually be a bit more work
than needed...

a simple custom-build middleware server
handling a pool for connecting to the DB
and handling turning any ResultSet into 3D objects
shouldn;t be too hard to implement

jeffs

Russ

2005-05-05, 7:27 pm

>According to >http://vrmlworks.crispen.org/e=AD>ai_faq/scripts.html,
VRML
>supports two kinds of Java scripting, >Java Scripting Authoring
>Interface
>(JSAI) and External Authoring >nterface (EAI). Only the EAI supports
>using events and fields that are not >part of the script node, so that

>is what we would probably need to >use. It says that EAI "is available
in
>the two major browsers".

>I don't know when this was written;

In an earlier age, when the world was
young.

>is it
>supported in IE and Firefox? Any other advice?

EAI is supported in I.E. with the
Microsoft Java Virtual machine only.
It does not work with Sun, at least
as far as mainstream VRML plugins are concerned.

Firefox: no. The other html browser
referred to is the Netscape 4* series
only.

Some VRML browsers have EAI that
do work with Sun Java:
Xj3D and FreeWRL, but their implementation is either as
a standalone application or in
applet form. Their EAI aspect does
not work in the conventional plugin
context.

I should add that both Cortona
and blaxxun Contact have an EAI
capability that uses Javascript rather
than Java, but I don't know if this
will work with Firefox.

tc
Russ Kinter