comp.lang.vrml FAQ
Answers
6. The VRML Cabal, politics, and miscellaneous whining
The VRML Consortium holds VRML in the firm grasp of its iron fist. When I wrote this recently, Gavin Bell, one of the originators of VRML, broke into fits of hysterical laughter.
In fact, control over VRML does rest with the VRML Consortium, and companies and individuals can join by simply paying the dues that keep the lights on, a part-time staff (poorly) paid, and the web site and mailing lists on which we rely on the air. Twice a year or more, there's a members' summit where any member can say anything, and the Board of Directors (BoD) and VRML Review Board (VRB) who are elected by the members generally pay close attention.
The whole organization and process is spelled out at the VRML Consortium's web site.
Technical control of the language is controlled by the Working Groups (WGs), who work on proposed changes and submit them to the VRB, whose motto is "just say no". Not that good ideas won't get into the spec. Not at all. But there has to be overwhelming support for them. Which brings us to our other player, the VRML community.
The VRML community as a whole meets on the majordomo list www-vrml each day and discusses the formation
of new WGs, the products of WGs, and, of course, the imminent death of both VRML and the Internet. Some kind souls
on www-vrml answer technical questions sometimes, but that's really what comp.lang.vrml
is for. But if you've got an idea about the language, or simply want to find out what's going on with VRML, you
should join www-vrml. Anybody can join.
Yes it would, and you have an opportunity to set things right. The VRML change process is completely open to anybody with an idea. Here's what you can do:
www-vrml, suggest the idea, and see if some
other folks there are interested in starting a WG with you.
That takes hard work and relies on both engineering skills and people skills. The people who set up the process figured that, if the spec was going to remain stable, changing it should be hard. Only the very best ideas survive this ordeal of fire. But maybe your idea is one of them.
Now let me tell you a story. I became very concerned a couple of years ago that VRML worlds looked wildly different
in the different VRML browsers. I made such a noise about it that people who actually understood color and lighting
got involved and realized that they'd better do whatever they could to take this issue out of my patently incompetent
hands. As a result, two Working Groups were formed, a standard for testing the color and lighting model was developed,
a Recommended Practice for a VRML Color Space is in preparation, hundreds of browser tests were built, and VRML
browsers now differ from one another only on the smaller details. I pitched in a little, serving as recorder for
the WGs, figuring that the WGs would have to explain what they were saying to non-experts. I claim no credit for
any of this, except that a good idea managed to light, however briefly, in my head, and I spoke up on www-vrml
and said something about it.
At any rate, if you're willing to come up with enough energy to complain, shouldn't you be willing to come up with a little more energy to help make it happen?
That's an easy one. PNG, which stands for Portable Network Graphics, is an international standard. The GIF format isn't. VRML is an international standard, so it just makes good sense to support other international standards. Besides, when you require a proprietary technology, you in effect levy a tax in the form of license fees on everyone who builds a VRML browser or application.
Now, in fact, all VRML browsers I know of support GIFs and not all of them support PNGs, but they don't have to support GIFs, and if they hope to pass the conformance tests, eventually they have to support PNGs.
It's the same for MPEG movies (a standard, and required) and AVIs and RealAudio movies (not standard, and not required).
VRML browsers may support non-required formats any way they like. In one case, this has led to unfortunate results:
one widely used VRML browser supports animated GIFs in ImageTexture nodes and another major VRML browser
supports them, but in MovieTexture nodes. The smart thing to do is not to use non-required file formats
in your VRML worlds.
No it isn't. Now please go away.
Actually, Intervista has a site that shows how VRML is being put to work in a number of technical and business applications. They've got a few products that are very definitely geared toward mainstream business, as does Oracle, and more are sure to be on the way. Visit them, and check out some of the things people are posting to this newsgroup. You can see for yourself.
VRML represents goodness, purity, and righteousness, while the others are the spawn of Satan and should be shunned by any right-thinking person.
In fact, Chrome (which is still under nondisclosure) allegedly has some very cool features which VRML can learn from. It isn't out yet.
MetaStream was just released and features progressive geometry, which is a very cool technology, but its browsing is limited only to examining static objects.
Java3D is a programming API, and the model it uses is that people will write programs which cause geometry to appear and do things. VRML is scene based, and its model is that scripts and the EAI manipulate objects in the scene. Horses for courses. Sometimes you want to do it one way; sometimes you want to do it the other way. Dave Nadeau has a set of slides that describe the differences and the choices. Sun and the VRML Consortium announced active cooperation some time ago and are working on how the two technologies can benefit one another.
Nobody understands XML. The VRML community are beginning to study it.
No it won't. We're doomed. Now please go away.
In fact, WGs are forming or have already formed to think about how VRML can best benefit from those technologies, and there's every reason to believe that people in Microsoft, Sun, and Intel (all members of the VRML Consortium) might think that some things in VRML will help their technologies as well.
Because if you're talking about 3D on the web, which a lot of people are finally starting to get excited about, the battle for what the standard will be is already over. We won.
Our strategy is simple: we will assimilate; resistance is futile.
You're absolutely right. They are. Now please go away.
Actually, there are features in those systems that we in the VRML community need to pay attention to. The folks who wrote the VRML spec didn't have all the good ideas that anyone can possibly have.
A number of the companies who have developed those technologies have been involved in VRML from the very beginning, and some of their efforts are coming before us now.
But the surest way to make VRML a dead standard is to stop learning from other people.
Oops. Wrong place to report this. While the engineers from these companies often read comp.lang.vrml, there is no guarantee that they'll read your question, and even if they do, they may forget to put your software problem report into their system.
Here are the places to report bugs:
There are two reasons. The first has to do with differences between platforms. PCs have, to name just one parameter,
a nominal gamma of 1.8 to 2.2, while Macs have a gamma of 1.0, SGIs can set their own gamma, and LCD laptops have
some other intensity transfer function. Very soon the Conformance Working Group of the Consortium will be recommending
a linear VRML color space and letting browsers, rendering libraries, and hardware perform the intensity transfer
function that will display that color space properly on monitors, printers, and other devices.
The second problem arises because VRML browsers can't write directly to each pixel. They have to use rendering
libraries like RenderWare, DirectDraw, and OpenGL if they're to have any hope of letting people move through VRML
worlds in real time. That means that browser makers have to figure out ad hoc what to tell the rendering
libraries so that the libraries will produce color and lighting that conforms to the VRML color and lighting model.
That's hard, but browser makers are cooperating through the Consortium's Conformance Working Group and they've
made a lot of progress. A year ago the color and lighting differences between VRML browsers were huge. Today, thanks
in part to this work and to Eric Haines's Pellucid,
a "perfect" implementation of the VRML color and lighting model, the differences show up in the more
subtle effects. When the conformance test suite is complete, we should be even closer to our goal of making VRML
worlds look the same everywhere.
![[line]](vrmlrule3.gif){kind=small}
Did I leave something out on one of these questions that you need to know? Let me know.
-- Bob Crispen
-- Thursday, June 11, 1998