VRML the Open Source Issues

Dateline: 3/10/99

So what is the story with the open source of the Cosmo code? CosmoPlayer is the preeminent browser for VRML97. PLATINUM is in the process of negotiating with the Web3D a license and process for turning over the source code to CosmoPlayer, CosmoWorlds, and PageFX.

First we need to clarify that the PLATINUM and the Web3D consortium are currently in negotiations about the exact status of the source. Regardless of the specific terms clearly the intent is to provide more and easier access to the source and that is a GOOD THING!

The entire world of "open source" is getting quite confusing. The FSF (Free Software Foundation) folks really started the ball rolling with the GNU (Gnu is Not UNIX) project which has been a resounding success. The big issue for FSF software is not only must the source be free but it forces software developers to release the source of any newly developed code. It codified it's policies via the now famous copyleft license arrangement. This of course has good and bad points. In addition there is now an OpenSource organization promoting the use of Open Source also a good thing! It's approach appears to be to allow commercial use and retention of some proprietary code. The OpenSource organization appears to be moving towards the use of OpenSource as a brand.

It's unclear if the access to the Cosmo source will be easy and/or wide enough. Platinum appears to want it both ways - here's the source but no commercial use. The current funding situation (or lack thereof) of the Web3D Consortium place severe limits on what it (the consortium) can do. My own personal preferences are to spread the source as wide as possible with as few restrictions as possible, of course I didn't spend millions of bucks for the source. The Consortium can control a master source tree and bless only those areas it wishes, preventing balkanization.

Neil Trevett President of the Web3D Consortium in a posting on 3/8/99 said:

Hi all,

The discussions between Web3D and Platinum continue positively. This is a brief note to update everyone on progress as it happens and to begin to clarify the direction in which we are headed. Everyone involved in the negotiations have been listening intently to the comments here on the list and I believe things are looking good so far.

Here are some guiding principles that are emerging as the consensus view - and are guiding the Consortium's negotiations:

1. VRML97 momentum It is essential that VRML97 thrive and grow. It is the standard that exists and is being used today and the Consortium should do everything it can to increase stability, conformance, deployment and visibility. Their is a strong view that these goals can be met by providing source into the community as early and as openly as possible.

2. X3D should be based on fast-track evolution - not revolution X3D will deliver many benefits and meet requirements that have emerged from our experience with VRML and the changing face of the web. However, X3D should resist re-engineering for its own sake, it should retain as much of the VRML infrastructure as possible consistent with meeting essential new requirements. Most importantly X3D needs to be timely - if it doesn't happen quickly it will lose market relevance.

3. X3D should strengthen VRML97 not weaken it This can be achieved by not gratuitously re-engineering VRML, striving for maximum backwards compatibility and rigorously defining the exact conformance mapping from VRML97 to X3D as early as possible. If planned and presented with care - X3D can be a positive boost to short-term VRML97 momentum as developers can have confidence in the longer term vehicle for their content as well as short term deployment on existing VRML97 browsers.

4. Open source is good - when coupled with a controlled source tree it is even better Open source enables widespread innovation and experimentation, unleashing powerful Darwinian dynamics. However a central source tree, and a license that encourages people to check improvements back into the tree compliments this community effort. It enables individuals' work to benefit everyone, and prevents balkenization.

5. The preferred license model is that used by W3C, which is "BSD like": public domain with a copyright clause -- you can do anything you want, you don't *have* to put back changes back into the "common source tree" - though it is encouraged. Commercially-friendly open source projects with a high degree of organized control over the source tree have tended to go this route (Apache, W3C).

6. Platinum (rightfully) wishes to retain the right to grant commercial licenses to companies who wish to use its source. Commercial use includes direct integration of the source into shipping commercial products. Platinum is willing to grant significant freedom to Web3D as to how the source is used non-commercially.

7. Some level of funding is going to be required by Web3D if it is going to be able to effectively resource short-term VRML activities and drive the specification of X3D. The status of the X3D design team, for example, will be undecided for another week or two as the dust settles from the Platinum layoffs. For example, it might be necessary for the Consortium to directly fund a full-time design team, or at a minimum to fund the infrastructure to enable the community to take over source-tree maintenance.

So - given all these desires and inputs, the direction we are headed in is:

1. Web3D is to be granted a license by Platinum, and is enabled in-turn to grant BSD-like, access to the source base.

2. Web3D immediately forks the source and enables two parallel tracks - each with an independent source tree:

a) The VRML maintenance track. An open license is immediately granted to the Community for all source, with a Web3D coordinated central source tree for checking improvements back in. The staffing for this source tree maintenance would ideally come from the Community with infrastructure setup by the Consortium. Anyone can download the source, for non-commercial purposes, at any time. For example, a company is deploying VRML content, and finds the browser has a bug that affects the integrity of its content. The company can download the current source, fix the bug, and send back the update for check-in. On the next source release, the bug is fixed for everyone. Commercial use on the other hand would include a company that wants to take parts of the source, and embed that source into its own product for sale. Such a company would need to purchase a commercial license from Platinum. Once that license is in place, the company can download and use the current source tree as it wishes (and feedback changes if it wishes also).

b) an X3D development track. Early access to this source tree would be available to Consortium members, though with frequent drops into the wider Community to keep everyone informed and involved. Coordination would be through a Consortium managed source tree. This is intended to keep things focused and to generate appropriate funding momentum. As soon as the standard is complete the reference source is made fully open just as track a). To clarify note that he X3D source goes pure open source when X3D is done and is available for both commercial and noncommercial use.

We expect to have a draft agreement from Platinum to enable this structure by the end of this week.

We will post more news as soon as we have it.

Regards, Neil

In addition to the activities of the Consortium another very interesting activity, led by Chris Marrin, the creation of the OVAL (Open VRML Advancement League). You can join the newly created discussion list and become one of the ragtag group of adventurous souls (Star Wars music fades in). This type of effort would be a joke were it not for the presence of Marrin (formerly the implementer of CosmoPlayer and developer of the EAI) and a fair number of respectable people. Clearly the urge for truly open free aavailableVRML source is burning brightly.

One Year Ago in Focus on Web3D VRML Content with a capital C

Previous Features