VRML: You may be building it but will they come?

A Keynote Address, planned for the 3D Design Conference

You have heard a lot about VRML and today you will get a really close-up detailed look. You probably know that VRML.. the Virtual Reality Modeling Language.. is making a bid for your collective 3D designer hearts and minds. Many of you already use the tools of the 3D trade every day. So what is this upstart, this VRML? Can VRML deliver on its promise of 3D everywhere, from the CD-ROM to the Internet to distributing the high end to the low end? Will 3D VRML worlds on the Internet attract millions of visitors? Should you invest in VRML now?

This presentation will not even attempt to answer these questions. There are plenty of other speakers today who will do a fine job of giving you the pieces to that puzzle. After you have heard them, you will be able to assemble your own picture.

The grim fact is that standards are universally praised and seldom adopted. This presentation will put VRML to the standards litmus test, comparing its evolution and current state with standards efforts of the past. I will also pose, and attempt to answer the question: what is the 3D killer app?

Been there, done that:

PostScript, lessons from the long road to the 2D standard

Many of you in this audience are beneficiaries of the success of a 2D graphical file format standard: Postscript. Who can deny the creative storm that was unleashed by the Macintosh, the Laserwriter, PageMaker, Illustrator, Photoshop and color imagesetters? Underlying all of these products and devices is PostScript.

The story of the adoption of the world's preeminent 2D graphical standard, has much to teach the proponants of VRML, who seek to dominate 3D content publishing. PostScript walked a very long road from its development by John Warnock, Chuck Geske and others at Xerox PARC to the formation and success of Adobe Systems Inc. Along the way, PostScript, a bulkier and slower page description language had to battle much more efficient but less feature rich typesetting languages. In the end, the lucky 'killer app' of desktop publishing, a fierce campaign to sell standard PostScript RIPs and a carefully guarded 'proprietary but open' standard put PostScript over the top.

PostScript vs. VRML

Consistent, commercially driven leadership by a single company and huge investment over a decade were key elements in the elevation of PostScript into a de-facto standard. VRML has a more distributed (but by some accounts not as open as it could be) design process and a leadership with varying levels of interest, commitment and investment.

Other key factors in PostScript's success was the creation (at great cost) of excellent tools for creative content developers, which hid them from PostScript and made the design process transparent. VRML has yet to develop a single comprehensive, and well designed file browser, let alone powerful visual development environments, none of which will be small development efforts.

Postscript began as a stripped down, simple subset of a much larger vision (a universal document description language) and was then gradually extended (adding better color, device support, portable formats and compression). Perhaps PostScript will never reach out to become an all-encompassing document architecture. There are now too many varied ways of handling documents. PostScript established its market on the strength of a simple, modest subset of its vision. Only then was it extended. VRML, on the other hand, was not yet established in any significant market for 3D before it was extended and is now being extended again. What is emerging is a very complex and bulky specification which attempts to wrap around an ever increasing design space. It is hard to imagine Adobe and PostScript surviving if file sizes and print times were ten times, rather than two or four times greater than its competitors.

Lastly, Postscript went head to head with its competitors, on a daily basis, for up to a decade before it finally vanquished most of them. Many customers found it too risky or costly to move from older systems that performed the job until the sheer force of features surrounding PostScript in their own field was undeniable. VRML has yet to go head to head with a single revenue generating competitive product. Where are VRML competitors to Doom or Nintendo 64, to WorldsAway, to Myst, to AutoCAD? Where is VRML even complementary to any of these successful products? What features of VRML capture or demonstrate the 'cool' features of these products. One could argue that PostScript grew into a market of its own making (it made everyone into a typesetter and graphic designer). However, PostScript made the grade in the traditional print marketplace alongside its success in desktop publishing. It was the marrying of the two (you could carry your PostScript file to the printers) that made PostScript a standard.

PostScript traveled a long road to become the 2D graphical standard. The creators of PostScript did not cut any corners, skip ahead prematurely or ignore the other travelers on the road. VRML is bypassing some important milestones on this road and risks arriving at the finish line alone.

Latin or Balkan:

Language and the fallacy of the single platform

The evolution of human languages is perhaps the most interesting of standards adoption processes. Pathways for acceptance of a language can be crudely divided into the Latin way or the Balkan way. In the Latin way, a commonly used ancient tongue is adopted widely as a second language (for personal interest) only to splinter into distinct but related languages (French, Italian, Romanian etc.). In the Balkan way, a procession of regimes attempt to impose a culture (and hence its language) over an unwilling population (Serb, Croat, Montenegrin, Macedonian, Romany). The resulting conflict seems interminable and can impoverish a region and cripple a culture.

From the Balkans rose no dominant world culture, and little science, art or literature. From the Latinate world rose much more, including one strange mutant: English. Cobbled together out of Latinate tongues and German, with its grammar simplified and 500 years of uninterrupted marketing, English now dominates the human experience. Above all, English is successful because of its simplicity and its porosity: it will absorb words and phrases from any language, it invites anyone in.

Computer Standards: Latin vs. Balkan

In the computer industry The Latin situation characterized the graphical user interface wars. The ancient ancestor of the medium (Xerox), now largely forgotten, set down the basic precepts and design. The immediate progeny, Apple, Commodore, Digital Research, Metaphor (and my own product of a decade ago, Elixir) created their own versions of the Xerox desktop. Then along came a persistent mutant. Cobbling together simple parts (DOS and 10,000 pieces of hardware and software) with dogged marketing, Microsoft became the lingua franca of the GUI universe. One could argue whether Windows is small or elegant (neither is English when you weigh its dictionaries), but its porosity and power to absorb is indisputable.

VRML runs the risk of following the Balkan way. Monolithic, complex, all encompassing and conceived exclusive of existing native languages, VRML may lack the basic criteria to be accepted as the lingua Franca of 3D. Will VRML ever perform fast enough for game environments, will it provide production values demanded by high end systems, will it support the user and cultural interfaces so successful in non-VRML social avatar worlds? Perhaps one language cannot possible do all these things (remember ADA?). Will VRML then become a jack of all trades but master of none? If this is the case, VRML will be relegated to its own small niche, leaving the separate worlds of 3D to duke it out.

For VRML to become a lingua Franca, a unifying force, it must be simple, open enough to adopt the features of other languages and, above all, driven by a single consistent commercially vested vision over a long period of time (like Rome, Britain, Adobe or Microsoft). English may be dominant, but all the Latinate (and even the Balkan) languages still exist. In human language, there is no common platform, only the occasional common tongues of translation and transaction, such as Latin and English. VRML must appeal to and change to serve the needs of many communities. Developing VRML for the VRML community has all the hallmarks of success of Esperanto.

Killer apps and production values:

creating worlds worth visiting

Let us assume VRML does become a lingua Franca. Through plug-ins, scripts and PROTOs, and a whole lot of good faith effort by individual developers and close coordination by the members of the VRML Consortium, acting effectively as a single vested commercial interest (or turned over to Microsoft), VRML emerges as the PostScript of 3D content publishing. Now what? If by this time, no other combination of 3D technologies has taken a decisive market share, VRML may be in a position, to live up to its early promise, to become the infrastructure of the new Cyberspace.

We might ask: what is the killer 3D app, what will make spaces in the new Cyberspace worth visiting? What makes any place worth visiting? One answer might seem obvious: places full of people are often visited. You visit a place mostly because of your affinity with the people and by the quality of the interactions you experience there. Other places are visited because they are not full of people. We go to the woods or hike into the back country of the Sierra because there is life there and it is not all human life. In third place are the spaces we visit because there are things there (not living, but perhaps once living), such as grocery stores, malls or cemeteries. There are some people who detest visiting places of things, but would certainly visit places of people or places of life. This is one reason why I put the places of things last.

Of course, bosses, teachers, court magistrates or other forces can induce us to visit any kind of place, whether we like it or not (yes, we will design these environments, too).

What does all this mean for the designers of successful 3D environments? It probably means that the 3D cybermalls will be soon boarded up if you design them to be navigated alone. It means also that we ought to pay close attention to how people interact in groups (a hard pill to swallow for computer geeks like me, easier for you creative folks with anthropology degrees). It means that we have to start thinking more about the designs of non-human 'living' things in digital space (bots, digital biota, A-lifes, A-I's) to make virtual worlds even marginally as interesting as the natural world. It means that object rendering doesn't matter as much as human communication and that polygons count less than the pirouette of virtual dancer.

Hold VRML up to your Standards

It also means that the proponents of VRML have to start focusing on a whole set of design priorities that are not getting much attention. As you attend each of the sessions today, ask hard questions of the presenters. If VRML cannot do something you deem essential in your 3D designs, challenge it! You don't have to accept VRML, it is not a standard, but merely a candidate. This day of VRML at the 3D design conference is one of the first times it has been presented outside of its community of believers.

For each VRML environment you are shown, ask yourself: "would I visit this place once, ten times, every day?". There are many 3D designed environments visited over and over again, in film, on TV, in books and magazines, in buildings and town squares (and even a few on the Internet). VRML has to pass the litmus test against at least one of these to be worth you spending your hours and dollars.

The Promise of VRML: You can make a Difference

The promise of VRML is that of a shared creative explosion orders of magnitude greater than set off by PostScript. A successful common 3D format holds out the possibility of becoming a planetary communications and management medium and a matrix for the survival and further evolution of life on Earth. These are lofty and noble goals. If VRML is not up to it, throw it away! If you feel that it has a shot, put it into the running by offering your insight, your talents, and your skills. And don't take 'no' for an answer!