Today's morning paper reports that yesterday --July 7, 1998--the price of shares in Yahoo rose over $26, to $199.25. At this price, the market value of entire company is more than $9 billion. Yahoo is therefore worth more than Union Carbide, one of the 30 stocks enshrined in the Dow Jones Industrial Average. The market value of America Online is currently in excess of $24 billion, almost equal to that of Sears, Roebuck, and more than that of seven of the stocks on the Dow. Amazon.com is now almost equal in value to Barnes and Noble and Borders--combined. Most of these high-flying Internet companies are showing little or no profit, but the market so greatly values their potential to dominate the economy of the near-future that it looks past the "growing pains" of the present.

Whether this is wise or foolish is not for me to say. But the implications in the Cosmo situation can hardly be ignored. In an environment that pays handsomely, even outrageously, for Internet technologies with promise, Silicon Graphics has been unable to find a buyer for Cosmo. SGI's troubles leave it in no position to continue to invest the future of VRML, and it appears that no one else believes in it enough to carry it forward.

For those of us who have been following the rapid developments in commercial VRML, this seems impossible. VRML was getting good. Really good. The last couple of columns on this site have been devoted to showcasing breakthrough work in VRML, and no one who has seen this content can doubt that the technology has matured and arrived. I have been unabashedly promoting VRML in these columns for quite a while because I was convinced, and still am convinced, that interactive 3D graphics on the Internet makes obvious commercial sense, and that VRML had proven itself as the appropriate technology for the job. Indeed, we were presently in the middle of a series of lessons on VRML, initiated by the excitement attending the release of Cosmo Worlds 2.0.

If Cosmo is shut down as planned, what will happen to this fabulous application? I have been teaching a class on VRML using Cosmo Worlds 2.0 as the development environment since the product was first released. The results that we are generating among a group of people with little or no previous exposure to VRML are simply remarkable. Cosmo Worlds is not only an outstanding way to develop in VRML--it's also the best way to learn it. The program is brilliantly conceived and executed, and thus it seems all the more incomprehensible that it would be simply abandoned.

One reads an "expert" for answers, not questions. It's far too early for answers, so let's try to make some intelligent observations, or at least refine our questions so that we can better assess the future of VRML and interactive 3D on the Internet.

It's wise to isolate the Internet element of the picture. Interactive 3D in the games and visualization markets is exploding. VRML is only one technology for delivering real-time, interactive 3-D applications and, being an open standard, is general in a way that makes it less competitive in specific markets. Thus a VRML browser is not as strong a games platform as most of the established games engines. And a number of firms offer much more powerful, and expensive, interactive visualization tools. But VRML's great advantage is its integration with Internet technology. Thus, for example, the VRML browser makes true 3D Internet gaming possible right now, and this has probably been the single biggest thrust in commercial VRML development.

So where does the departure of Cosmo leave those of us with a particular interest in 3D content on the Web?