David writes:
>This confirms my suspicion that GeomView is not a VRML compliant browser.
Correct. Geomview is not currently a VRML compliant browser. *After*
we release the WebOOGL 2.0 toolkit RSN, it will be "VRML
quasi-compliant". That is, it will read and display VRML files,
ignoring nodes that Geomview doesn't support (most notably textures).
Justin writes:
>We've been through the standards process. [...] The fact is that OOGL
>is *not* VRML. [...] There was a clear consensus (albeit certainly not
>a unanimous one), as long as six months ago, that Open Inventor was
>the way we were going, and it is the way we *have* gone.
Yes! God forbid another standards debate on 1.0. The VRML 1.0 spec
*is* an "extended subset" of the Open Inventor file format, *not* the
WebOOGL file format. I'll repeat what I said in person at the Monterey
BOF: I fully support the VRML spec as it stands and think that the
choice of Open Inventor file format as a base was wise. For those new
to the list, I'll point out that I say this as the primary author of
the second place proposal. See Brian Behlendorf's message from 18 Aug
1994 for the details of the voting results (message ID
<[email protected]>).
My intent with the WebOOGL proposal was
1. to build a working and distributable prototype system around
Geomview with a minimal amount of work and
2. to influence the VRML debate with our philosophy of 3D viewer
systems.
The former was accomplished, the latter will really get hashed out
during the VRML 2.0 hubub. One of the chief strengths of Geomview that
I'd like people to focus on is GCL (Geomview Command Language). It's
the Lisp-like command langugage that gives Geomview its strength and
flexibility. GCL allows Geomview to be used as a display by separate
running programs or to have its functionality extended by people
completely unfamiliar with its internal stucture. It's because of GCL
that I could get the WebOOGL prototype working via perl scripts in a
matter of days without touching the source code to Geomview itself.
While the original GCL design decisions were not made with the Web in
mind, I think enough of the issues are similar that I'll add my
hindsight about its design to the VRML 2.0 debate.
A clarification on the relationship between GCL and OOGL: a subset of
GCL is OOGL, which is technically just the geometry/appearance data
formats. This perhaps subtle distinction is not generally made on this
list, most people have been using OOGL to refer to all of GCL. Roughly
speaking, OOGL corresponds to VRML 1.0 and GCL corresponds to one
possibility for VRML 2.0. Much of OOGL is historical, which is of
course not generally considered a compliment. While OOGL/WebOOGL is
indeed a valid 3D format in its own right that we will continue to use
as our native format, I fully expect that most people will use VRML as
a common interchange format.
Again repeating a comment from the Monterey BOF, I encourage everybody
to write converters to and from their favorite formats and retrofit
their tools to accept VRML as an input/output format. Sure, VRML 1.0
isn't perfect, but it's a good first step. Also, one of its heretofore
unsung virtues is as an interchange format for plain 3D data with no
embedded URLs.
Linas write:
>Check out the Geometry Center at univeristy of minnesota.
>They have had a functional vrml browser for over half a year now.
So the upshot of all this is that I'd call it an "integrated 3D Web"
not "vrml" browser since I'm asserting that VRML is the draft spec not
just a catchall term for 3D file formats with URLs. But despite me and
everybody else jumping down your throat on the definition of VRML, let
me say thanks for pointing out that we've had something up and running
long before this month's hullabaloo :)
Mitra writes:
>Any chance of a URL for this, I couldnt find it there.
The WebOOGL URL is http://www.geom.umn.edu/locate/weboogl/
End of diatribe, time to do my taxes.
Tamara Munzner The Geometry Center (612) 626-8325
[email protected] http://www.geom.umn.edu/~munzner