Re: Statement from SGI about Inventor

Brian Behlendorf (
Mon, 27 Jun 1994 15:25:23 -0700 (PDT)

On Fri, 24 Jun 1994, David Cake wrote:
(quoting Gavin Bell)
> >While we reserve the right to make changes to the core Inventor
> >format, we do not expect to make any such changes in the near future.
> >Every Inventor file is required to have a header line identifying the
> >version of the format; we will not make changes to the core format
> >without incrementing the format version number in the header.
> It is quite likely that if VRML takes off, that there might be more VRML
> code out there than plain Inventor code after a while. Look at the current
> explosive growth of WWW pages. In other words Inventor compatibility is not
> necessary to keep VRML alive, only to kick start it. The two will probably
> be fairly different by the time VRML 1.0 is finished (presuming that we
> start with Inventor), and if SGI change the Inventor format in a strange
> incompatible way after that, there is no necessity for the two to stay in
> synch. How do SGI feel about this?

In other words... is SGI worried that the elements of Inventor we define
as VRML become so widely used that they will be hamstrung against being
able to change it in the future? Seeing as how they want it used widely
in the first place, I don't see this as a potential problem. Consider
HTML and SGML - HTML is defined (at least initially was, and is being
brought back into compliance) as a Document Type Definition under SGML.
Almost anything those working on enhancing HTML want to do can be done
under SGML, as SGML defines syntax and structure, not presentation.
There are several companies with SGML products, and they are definitely
benefitting from increased interest in HTML authoring and conversion
tools. Likewise, it seems any extra data types we create could be
defined in the Inventor language, and thus stay "in synch". SGI has
a committment to their customers and third-party vendors to not shift
low-level details without warning and dealing with legacy issues anyways.

> I guess we have our answer. Now we might actually have to look at
> Inventor on technical grounds.

Definitely. Last week I bought "The Inventor Mentor" by Josie Wernecke
(ISBN 0-201-62495-8) and I'm reading up on it. I'll try and track down
more info we can put up on the web site...