Michael Frank Deering: HoloWeb

Date of Proposal: February 1996


HoloWeb was a last minute submission to the VRML II competition.

HoloWeb was a reaction to the almost total lack of progress in improvement in the main proposed VRML II design over the VRML I design. I had sat in numerous meetings with the people involved in this efforts (including several SGI engineers), somehow every needed fix that everyone agreed to would mysteriously be retracted at the next meeting, with a vague explanation that “an important customer wants it the old way”. (I never understood where these decisions were coming from, though I have my theories.) About the only change that stuck was taking side-effects out of the traversal of the scene graph.

So at the 11th hour (literally a few days before the VRML II proposal deadline) as a protest vote I submitted a radically alternate proposal for VRML II based on compressed geometry and many of the concepts from HoloSketch (and that would eventually make it into Java 3D). I didn’t really think that it would be a contender in the competition, but I wanted to make a statement about all the things that I thought was wrong with VRML by suggesting specific alternatives. Microsoft and Apple also entered last minute VRML II proposals (likely) for similar reasons. HoloWeb only garnered 5 votes (I didn’t even vote myself), but people do remember, so it made a difference. As of 2004 (8 years later), VRML’s successor, the Web3D consortium, is in the process of adding compressed geometry to their specification.


HoloWeb’s main concept was to organize a 3D world-wide-web around 3D virtual real-estate. Using 768-bit high resolution coordinates, every web site could have its own virtual 3D space for its site and all its pages. The concept looked a lot like the universe imagined by Neil Stephenson’s Snow Crash. (This was one of the things a lot of people liked about it.) However, the concepts were based on ideas I had from before Snow Crash was published (sorry Neil).


Below is a pdf of original submission. However, currently the file below is in inverse page order! This copy must have been meant for old printers that needed files printed in inverse order to stack right. I’ll eventually get a fixed one here.

pdf (247 KB)