High Programmer > Alan De Smet > Hypercosm and Web 3D

Hypercosm and Web 3D

by Alan De Smet

[Graphic: Hypercosm Logo] [Photo: Hypercosm Studio box]

Hypercosm Screenshots

[Graphic: DNA applet screenshot]
How DNA replicates.
[Graphic: Gaits of motion screenshot]
How animals move.
[Graphic: Mars Global Surveyor screenshot]
Mars Global Surveyor (for New York Times on the Web)
[Graphic: Air Hockey screenshot] [Graphic: HyperBubble screenshot]
[Graphic: MoshBall screenshot] [Graphic: Similarity screenshot]
Some games.

Hypercosm has regrettably gone out of business, making it hard to link to. This page exists to remember what Hypercosm once was.

Update: Hypercosm appears to have returned from the grave! They're using Orbitec's address. My best guess is that Orbitec has decided to revive the old brand. I don't know much about it.

New! You can now download Hypercosm and see some Hypercosm applets.

What was Hypercosm?

Hypercosm produced a specialized programming language called OMAR (Object Modelling And Rendering) suitable for describing 3-D content. Hypercosm made available a web browser plugin for displaying this 3-D content. Unlike many other plugins, the Hypercosm plugin was available for Windows, MacOS, and Linux systems.

Hypercosm was a great place to work and the technology, while imperfect, had a great deal of promise. For all of the weaknesses we had some great applets. We had an impressive 3-D cut-away lock that showed how pin and tumbler locks function. (Almost every lock you deal with, including your car, home, and office is a pin and tumbler lock.) There was a large selection of applets for the National Science Center showing how a variety things worked including gravity (with a cannon firing game), Global Positioning Systems (you could fly a helicopter around the world and see satellites report your distance and how you triangulated your location), and a host of others. New: The NSC applets are still online! check out the National Science Center Hypercosm applets. For Space Explorers Marslink and Moonlink were developed, allowing students to plan and execute simulated space exploration missions. As part of Moonlink students could participate as a team on multiple machines acting as ground-control to launch a rocket. A host of games were developed. It's a shame that all of this content is now lost.

Since Hypercosm's demise the player has disappeared from the Internet. I have recently made the last available player and development tools available online. Download Hypercosm Player or Hypercosm Studio. What screenshots I was able to scrounge from the web are to the right. I've collected some applets here.

Around spring of 2001 it became clear that Hypercosm was going to fold. "Francis [Hypercosm's CFO] admitted during dinner that Hypercosm is burning through $250,000 a month and has until the day after New Years to live. They need another $1 million in angel backing." (from the article "Dining among the angels...", I-Street, December 2000. By late spring of 2001 all staff beyond minimal management was laid off. The managment team attempted to sell the assets, I don't know if any were purchased.

In 2003 The Capital Times ran a belated article on Hypercosm's fall.

Thanks to the Internet Archive you can still see one of the last revisions of the Hypecosm web site. Because of the nature of the Internet Archive, many images will not load successfully, but you can still read much of the web page. If you're interested, there are over a dozen different versions of the Hypercosm web site available

Mark Taylor, once Hypercosm's web designer has put together a bunch of links to various ex-Hypercosm staff. This is useful if you're looking to see where the staff ended up, but is not so good if you're looking for someone to ask about Hypercosm's technology or assets. At the moment I don't have any links or other contact information for the ex-Hypercosm management team and it doesn't appear Mark does either.

Hypercosm's History

I've quoted a number of people involved in this story. All quotes are used with permission. While I've quoted large blocks of text, I've occasionally reorganized their words to match the flow of this document. I've also made minor editing adjustments, reformatted the text for the web, and added links to other sites. If someone sounds a bit disjointed or confused, is placing odd emphasis on something, or a link implies something, please understand that it is probably my fault as the editor, not the person I am quoting.

1989-1993 - Abe's Ray Tracer

Around 1989, Abe Megahed developed ART, Abe's Ray Tracer, at the Undergraduate Projects Lab (UPL) at the University of Wisconsin. The earliest reference I can find online is an October 1991 post from an ART user. The earliest official announcement I can find is Mark Spychalla's announcement of the availability of xdart for HP workstations. Neil G. wrote in to say that when he joined the UPL in the summer of 1989, ART already existed in a recognizable state.

Possible commercialization was already a thought, in March of 1991 Abe and two other students sought outside experience in marketing "a rendering program." Presumably the program in question was ART.

1993-1996 - Microcosm

After ART Abe and others developed Microcosm. While I don't know if any ART code was actually part of Microcosm, Microcosm is clearly the successor to ART. It supported ray tracing and interactive rendering. While Microcosm still supported the ART language, a new language designed to be easier to use, the Simulation / Modeling Programming Language (SMPL), was added. this 1994 post is the first reference I can find to Microcosm on the web. In the post features include "Real-Time Soft Shadows, Reflections, and Transparency / Refraction". This would remain a relatively unusual feature in similar packages and was a highlight of Hypercosm later.

Apparently around 1993 Abe and Mark formed the Cosmic Software Corporation and tried to take Microcosm commercial, as seen in this post to Usenet. Interestingly, in that post they said, "Microcosm was inspired by the type of programming which was encouraged by the days of programming in interpreted BASIC on the Apple II. Since those days, computers have become faster and more powerful, but they have also gotten more complex and difficult to use and especially to program. This has driven the computer away from schools as an educational tool in the sciences and relegated it to run canned programs such as word processors instead of running the students' programs. Microcosm is meant to change all that." This philosophy was still held by Abe, Mark, and Perry at Hypercosm.

1996-1999 - Megahedron

[Graphic: Just Say No to Ray-Traced Reflective Spheres over Checkerboards]
The "Just say 'no' to ray-traced balls over checkerboards" logo

Later, perhaps around 1996, some agreement was reached with Syndesis Corporation to market Microcosm under the new name Megahedron. (Apparently there were possible trademark conflicts with Microcosm.) The Internet Archive has a copy of the Megahedron web page at the time. On that page you can see some classic Microcosm imagery, including the No Reflective Spheres Over Checkboards logo. Here is a 1996 announcement of Megahedron on Usenet.

John Foust of Syndesis wrote in with some more information. I've reformatted his message and added some links.

Mark Spychalla worked for Syndesis from late 1995 until January 1997. I met him because I saw their postings about Microcosm in January 1995. Mark interviewed in August, after Abe read an article I published in Byte in July about 3D APIs.

After a while, I thought I'd help them develop it for a few more platforms and publish it in a wholehearted way. Sadly, it met with a very cool reception. We had a hard time giving it away even to the most nerdy folks at SIGGRAPH conventions.

-- John Foust of Syndesis by personal email

1999 - Hypercosm Forms

[Graphic: My old business card]
My old business card
[Photograph: Hypercosm manuals]
Hypercosm manuals, Hypercosm Studio CD, and mini-CD giveaway.

Around Spring of 1999 Hypercosm was formed by Abe and Perry Kivolowitz. Perry Kivolowitz founded Elastic Reality, the company that popularized morphing technology (Perry actually got an Academy Award for Technical Achievement in 1996 for that technology). The idea was to use modern computer power and hardware acceleration to make Microcosm real time and suitable for use in a web browser. 3-D seemed the natural direction for the web to go. Various groups tried to push VRML as a solution, but it failed to capture a significant following. Microcosm seemed like a natural fit to the remaining hole in the web. The company quickly formed and scrambled to produce their first version in just a month or so to unveil it at a conference that spring.

Unfortunately all of the legal issues of Syndesis's claim to Microcosm weren't worked out until September. Syndesis' current Megahedron web page hints at a payoff.

2000-2001 - Hypercosm Tries Business Plans

Hypercosm tried a number of techniques to make money. The first plan was to charge for development tools and provide a free web player (similar to Macromedia's sales of the Flash development tools). This failed miserably. Hypercosm also tried giving the development tools away but charging to post applets on the web using a key system to restrict access to applets. That failed. For those people who didn't want to buy a key, we tried selling advertisements that played before content without a key could be viewed. Still no one was interested in buying keys and no one wanted the ads. This also failed. Throughout Hypercosm developed custom applets for a variety of companies, this proved to be the only steady income source and it didn't cover the overall operating expenses. Ultimately there wasn't alot of demand for 3-D on the web. Useful and good applets were prized, but the market didn't care if they were 2-D or 3-D. Looking at the web today (2003), this still looks to be the case. Perhaps 3-D on the web is a doomed proposition for the immediate future.

Hypercosm, for all intents and purposes, ended in May of 2001 when the majority of the staff were laid off. Abe Megahed has some comments on the last year or so of Hypercosm's real life.

At Hypercosm, we became pretty focused upon the consumer oriented "customer care" application which I still think will eventually become the killer application for this sort of technology. However, it became clear to all of us towards the end that the world was not going to change quickly enough for this to become a viable business for Hypercosm. It also became clear that the funding was mostly dried up and that we were not likely to get funded by venture capitalists or other investors. Given that as the premise, there were really only two options available to us: (1) find a solid company to acquire Hypercosm so it would continue until the climate was more favorable for success (2) find a different market application that could support the business at some level.

In terms of acquisition, our best bet at the time was Adobe. From a technical perspective, we figured that Hypercosm could integrate with their current 3D product offering (Adobe Atmosphere) and our technology seemed to fit their company as a 3D analog to postscript. Also, I think we hit it off well with the people at Adobe and I think they had a decent respect for us. The technical lead at Adobe in the 3D area was a guy named Michael Kaplan and I knew of his work and publications. He and I became friends and I visited his house a few times. We even had a deal for a time that Hypercosm would get distributed with each copy of Adobe's web authoring tool, GoLive. The problem was that Adobe got into bed with the company Metastream, which is notoriously cutthroat. From that moment on the relationship with Hypercosm cooled. Adobe also reneged on our GoLive distribution agreement. So, under those circumstances, the prospect of acquisition from Adobe dissipated as quickly as it had begun. Other prospects such as Macromedia and AutoDesk (Discreet) were never really options. At that point in time, companies were pulling back on expansion and we didn't have a tight personal relationship with the people from any of these other companies. So, if you were ever wondering what we were doing in our expensive offices out in San Jose, this was a large part of it. It's amazing how much of business is still a matter of personal relationships and we always had one hand tied behind our back because of our location in Wisconsin.

The other strategy was to try to find a different application that could support the business at some level. I looked at three main "vertical markets": Architecture, Aerospace, and Industrial Applications. We looked at the architecture market by doing a prototype of the Overture Center in Madison and showing it to the Overture executives and the architects from Cesar Pelli. The Overture guys liked it but the architects were still more comfortable with the hand drawn renderings. Given their conservative reaction, that avenue was pretty much closed. I'm still amazed at how little these tools are used in architecture. I don't know why every architecture firm doesn't have a digital projector and some interactive 3D software. Anyway, similar story with the industrial applications that we looked at - they just were not jumping on it. The last area that seemed to show promise was Aerospace. In this case, we were lucky to have started out Hypercosm's operations in the office space across from one of the only real aerospace companies in the state: Orbitec. Way back in late 1999, I showed our venerable HP printer demo to Tom Crabb, a founder of Orbitec and his reaction was "This is great: NASA doesn't have anything like this!" So, together, we drafted an SBIR proposal to NASA and sent it off. Later, in the spring of 2000, I drafted 2 more NASA proposals, but these were both soundly rejected. In retrospect, I was asking for too much money up front because of our desperate situation and there's no way that NASA would have awarded such a contract to an unknown right off the bat. There was simply no way to save Hypercosm in the form that it was in.

-- Abe Megahed by personal email

An entertaining side-tale: Hypercosm's original office was in the Space Center on Fourier Drive in Madison, Wisconsin. The building was presumably named the Space Center because the company with the largest space usage was Orbitec, a company that focused on space exploration. While working in the building, every once in a while we would feel a faint, but distinct vibration. People from California say it felt like a very minor earthquake. As it happens, Orbitec apparently occasionally tested rocket engines in the basement, and we might have been feeling those tests.

Hypercosm's assets were acquired by the State of Wisconsin and the Wilshire investment group from Milwaukee, our only secured creditors. The State was secured because the money that we took from them ($500,000) was technically a low interest loan rather than an investment. The Wilshire Group got in at the very end and was able to get a secured position with an investment of only $250,000 because of our desperate situation. It's not exactly fair since we had some much larger investors who were unsecured and lost everything. I guess that's how it goes. I tried to find a way to get something (stock in the new venture) for the unsecured investors even as a token offering, but it just wasn't feasible. I had to tell a few investors that they were out of luck which is no fun at all. Incidentally, the Wilshire group had other reasons for putting the money in so they didn't exactly lose out. They were putting money in to Hypercosm mainly in order to be eligible to invest CapCo money from the State of Wisconsin.

So, what exactly are the Hypercosm assets? Well, it includes the software, content, web sites, trademarks, and interestingly enough, the patent that we filed way back in 1999. A few months ago, my wife, Doren, was bored at work and so she did a Google search on me. One of the things that came up was a notification of the award of our patent. If she hadn't done this, I would have never been informed either by the USPTO or by our patent attorneys at Quarles and Brady. Nice, huh? So, technically, we have a patent on what is essentially the idea of 3D postscript - combining an interpreted language with 3D graphics operations.

Very, very few inventions are truly original and without precedent of some kind and probably no inventions of any significance are without precedent. I hope someday to get a patent on toast. It remains to be seen what the significance of this patent is. It could be worth something.

-- Abe Megahed by personal email

The patent in question is US Parent 6,426,748 ("Method and apparatus for data compression for three-dimensional graphics.") The patent describes compressing 3D graphics by describing the scene in terms of a virtual machine tuned to displaying 3D graphics. This is exactly what Hypercosm did. In a sense it's a "3D PostScript".

This makes three patents Abe has been granted. The other two are US Parent 5,497,453 ("Method and apparatus for detecting and visualizing interferences between solids." This is about cutting cross-sections from arbitrary 3D objects, a technique present in Hypecosm.) and US Patent 5,428,716 ("Solid-clip methodology and architecture for clipping solid models and displaying cross-sections using depth-buffers," again related to cutting cross-sections of 3D objects.)

2002-2003 - Hypercosm officially closes

Hypercosm Inc. does not exist any more as a corporation. For a long time, it existed in a sort of zombie state until I finally got around to filing the paperwork to formally dissolve the company (only about 3 months ago). I only recently received notification that the paperwork had gone through to finally dissolve it.

-- Abe Megahed by personal email

That is, Abe filed the paperwork in fall of 2002 and the corporation was formally dissolved in spring of 2003.

2001- - Hypercosm Reborn

In early spring of 2003, I exchanged some email with Abe Megahed, one of Hypercosm's founders. It turns out that Hypercosm finally found a niche in supporting NASA! The Hypercosm technology has already been used for training purposes by NASA and may be used for some emergency training as NASA attempts to scavange scientific projects harmed by the Columbia disaster. Orbitec now owns the rights to the Hypercosm technology. Here is Abe's take on the current situation:

Orbitec Logo]

Finally, on September 11, 2001, nearly 6 months after Hypercosm had already gone out of business we finally heard that our original NASA proposal had been accepted. The Phase 1 project was small - Phase 1's are typically tiny - only about $60,000 to $100,000 so this wouldn't have saved Hypercosm anyway. This Phase 1 was a 6 month project to create a set of virtual 3D simulation based training materials based upon Oribtec's scientific payload, the Biomass Production System (or BPS) which was used to grow plants on board the International Space Station in the spring of 2002. After completing the Phase 1 project (see photo attachment), we began to show it around to crew members and crew trainers at the NASA centers, mainly the Johnson Space Flight Center in Houston and the Marshall Space Flight Center in Huntsville, Alabama. The response has been uniformly positive. In the fall of 2002, we were finally awarded a sizeable Phase 2 project to continue this work. This is what I'm currently working on and it will be at least a two year project. After a long a protracted discussion with Hypercosm's creditors, we have finally managed to transfer the rights to the Hypercosm technology over to Orbitec. Also, I've hired on one additional system programmer to help me and we are looking for a content specialist to help with some of the simulation work that is starting to come in. So, the Hypercosm saga continues...

You may be wondering what the impact of space shuttle Columbia's destruction may have on the future of this work. The future remains to be seen, but in the near term it looks like the current situation may actually increase the need for our simulations in some aspects. As you have probably heard in the news, we have 3 astronauts on board the International Space Station who will likely be there longer than initially planned because of the Columbia accident. NASA is looking for ways to get scientific experiments up to them on the Russian Soyuz. The problem is that the crew will not have been trained on these payloads. The solution: Graphical just-in-time training techniques using 3D simulation. Hypercosm simulations can make operations clear enough that a person who has never even seen a piece of equipment before can successfully operate it. If all goes well, I hope to be able to send some Hypercosm simulations to the Space Station soon. It would be very satisfying to be able to help NASA out in its current situation and for something positive like this to come from the ashes of Hypercosm and all of the wasted time, effort and money that went into it.

-- Abe Megahed by personal email

Hypercosm's competition

While working at Hypercosm, I generated a list of our competitors so I could check on them occasionally. Metastream, WildTanget, and Cult3D were only ones we really took seriously. At this least this might represent an entertaining view of 3D web technology in 2000.

The only company that seems to have survived in precisely Hypercosm's niche (general web 3D) is Cult3D, but even they seem to have failed to take the web by storm. No one seems to have atained the level of flexibility and interaction that Hypercosm assumed was necessary.


Metastream was a direct competitor to Hypercosm. They has a very slick 3-D plugin. It looked really good, but didn't appear to support any interaction beyond spinning a 3D object. They didn't have Macintosh or Linux versions. Apparently they did as well as Hypercosm, I'm having problems finding a concrete web site for the company. If you're interested 3D Gate has a September 1999 review of Metastream. Metastream was a product of Viewpoint and now appears to have been quietly rebranded as the Viewpoint Media Player, but even that is hard to find on their web page.


Claim to be "Driving streaming 3D technology." Founded in part by Alex St. John, an ex-Microsoft employee. (At the time his Microsoft connection was heavily touted as an advantage for unclear reasons.) They had a strong focus on game deployment. It used at the time native binaries, creating a large security risk and ensuring that there would never be Linux or Mac support.

Check out WildTangent's current web site. They have exclusively focused on games at this point. You might also find an archived copy of WildTangent's web site at the time Hypercosm was competing with them interesting. (Thanks to the Internet Archive for the historic link.)


Another surviver, you can see Cult3D's current web site, or through the Internet Archive you can see the historic Cult3D web site. It's been a while, but I remember Cult3D as having gorgeous graphics and a lightening fast software renderer (important in 2000 when many computers didn't have hardware acceleration). They could also overlay 3D content over normal HTML by taking advantage of some undocumented behavior in web browsers. Again, it supported only minimal interaction was basically a way to spin 3D objects.

Graphic Gems

Claimed to be "The 3D Web Solution". Their web site is long gone. The prior existence of the Graphics Gems series of programming books makes it difficult to find any references to them. I recall that they supported spinning 3D objects, and not much else.


Out of business, they have kept up the Unrealty home page. Based on the Unreal game engine, Unrealty focused on providing very realistic 3D "locales" to explore. As is clear from the name, the goal was to allow one to see remote locations, perhaps a realtor might use it to sell a house, or an architect to sell a shopping complex design. It was available for Windows and MacOS, a Linux version was in the works but cancelled with the rest of the project.


Research work being done at Carnegie Mellon University, Alice continues. The current Alice web site describes the new version. At the time Hypercosm was competing with the older Alice99. It was and is Windows only, but they report that they are considering a port to MacOS and Linux.


Hollywood3D was purchased by WildTangent. A Historic Hollywood3D web site is available courtesy of the Internet Archive.

Contact webmaster - © 2003 Alan De Smet