Re: Open Inventor *as* VRML

Linas Vepstas (linas@innerdoor.austin.ibm.com)
Fri, 9 Sep 1994 15:50:19 -0500


> From: Mark Pesce <mpesce@hyperreal.com>
> Subject: Open Inventor *as* VRML
> To: www-vrml@wired.com
> Date: Fri, 9 Sep 1994 12:04:26 -0700 (PDT)
> Cc: brian@wird.com, mpesce@taz.hyperreal.com (Mark Pesce),
> rikk@sgi.com[3~[3~[3~[3~[3~[3~[3~[3~[3~[3~[3~OD
>
> VRML List Members:
>
> [ ... ]
>
> and asked people who have a considerable background in graphics and/or
> languages, parsers and compilers to evaluate OI, then report their findings.
> The results came back *uniformly* positive;

THIS IS NOT TRUE. I've posted a number of notes to this group
bashing Inventor. I'll do some more below.

There have been numerous publications in the trade that have
negative comments about the technical implementation of Inventor.
Since I've heard even more negative things in private conversation,
I can only conclude that you interviewed a narrow segment of the
population.

See: (Why can't I find my copies when I need them?)

1) "A Declarative Representation System for Dynamic Visualization"
Mark Alan Tarlton, dissertation, Univeristy of Texas, 1993

2) Several arcticles emanating for Brown University group, e.g.
the 1993 Course on Object Oriented Toolkits at SIGGRAPH.

3) "GROOP: An object-oriented tookit for animated 3D graphics"
Larry Koved, Wayne L. Wooten, OOPSLA '93

> [ ... ]
>
> Nevertheless, we don't have much time at all. We will be presenting a
> specification document at WWWF '94, just *six weeks* from now. We can have
> a specification done by then, we can present it to the community; but if we
> argue on whether to use BEGIN or {, it'll be the millennium before we get a
> specification written.

Having participated in good faith in the VRML discussions, I wouldn't
be too happy if you went off and published a paper that said "VRML
is Inventor." I have no problem your saying "Inventor Integrated
with WWW"

> [ ... ]
>
> 2) Commercial arguments for using OI
>
> [ ... ]
>
> A commercially accepted standard makes sense, technically AND economically.

An unbiased presentation usually presents the downside along with the
upside. Since you seem not to mention any downsides, let me:

I am seriously concerned that adpoting Inventor in its current state
would be a serious impediment to future enhancement, upgradagbility,
and tunability of VRML viewers. The Inventor file format resembles
the Inventor API, which, in turn, is uncomfortably close to the
OpenGL API.

I am concerned that once you agree to read Inventor file format,
you will discover that you will have to use Inventor source code
to visualize the stuff. You will not be able to create a radically
different visualization library underneath the veneer. And if you
can't do that, you can't performance tune, and therefore, you
can't market your viewer as "faster" or "better". Your only choice
is to license your code from SGI, and you'll learn to love it, warts
and all.

I beleive that similar remarks apply regarding extensions &
enhancements. I've even gotten Gavin Bell to admit that its
hard to add a new primitive to the system (I apologize that
I have to quote you, Gavin).

(This should not be new news ...
Take, for example, Motif: it is uneconomical to do anything other
than to license the code and recompile it, and attempt to sell that.
Does this benefit the customer? Hardly -- there is very little
variety of features & performance to pick & choose from. All this
because its technically very difficult to improve Motif.)

> [ ... ]
>
> While it may seem that SGI is giving "something for nothing", remember that
> the widespread acceptance of tools that use OI will only increase the market
> demand for the machines natively designed to run OI, that is, SGI
> workstations. SGI is a hardware company, and this will increase the demand
> for SGI hardware, if it's anywhere near successful.

I've heard computer industry professionals use the same words as drug
dealers: "The first bag is free." (None of these people work at IBM,
and I'd get sued for libel if I told you who said this).

> Final thoughts:
>
> OI and VRML will not be quite the same thing, unless OI extends itself, by
> definition, to include the new nodes which we must add to give it VRML
> functionality.

I would be very open to basing a VRML that was a stripped-down version
of the Inventor file format. There is a lot of goodness in Inventor,
along with a lot of junk.

> ecology of the greater Web environment.

And if you don't own an SGI box, you must not be on the Web ??????

> We have a fantastic opportunity here, and a great offer from an outstanding
> group of people at SGI. My advice is that we quickly move to adopt OI as
> the basis for VRML, and begin the *real* work of implementing the tools and
> technologies to concretize our visions.

YOU have a fantastic opportunity here. YOUR company just might make MILLIONS.
Masking this realization as a non-partisan "contribution to the net" which
"we" should endorse is disengenous, embarassing, and a disservice to the
community.

> Mark Pesce
> VRML List Moderator
> President
> Labyrinth Group

--linas

Linas Vepstas
Graphics Architecture,
GL and OpenGL on AIX/6000 and OS/2
IBM Corporation