In the heady and dangerous world of computer graphics, there are the suits, and then there are a joyous group of idiots who have mistaken product announcements for actual business opportunity.

BitBoys Oy! (emphasis on Oy! - now there's a trademark in ridicule if I ever saw one) is a Finnish entity (I have trouble naming that a company) that made itself a granit-embedded reputation starting around 1998.

That year was the Year Of The Voodoo. 3Dfx was making a killing in the field and these jokers decided that they would also make a great 3D graphics chip. After all, the market was new, what was there to lose ?

So, they made a product announcement : the Pyramid 3D. A number of important buzzwords were included, such as optics-based architecture (??), the mandatory photorealistic detailing and even programmable pixel pipeline. Of course, no product announcement would be complete without a number for the fans to drool upon : 1 million shaded, textured z-buffered 25-pixel triangles per second. A final note indicates that the chip will be compatible with VGA - gee, how nice.

But that did not make enough noise, apparently, and the "project" was shelved for another great announcement which heralded the coming of the Glaze 3D. This was a new 3D chip boasting FOUR TIMES the power of the Voodoo 2 (which just happened to be the most powerful chip on the market at that time). This chip would have the works : aniso filtering, SLI performance and TV OUT (remember - we're in June 1998 !). Although, if you read the interview, they specifically state that (they) have not yet decided what method (they will) choose for Anisotropic filtering. Okay, so they are in the planning stages - fair enough.

Yet, in May of that year, the planning stages were obviously behind them, following this press release. Every buzz word is there. 400 million this, 200 million that, and fastest ever whatever. The title clearly indicates that the time is NOW.

So, in May they are boasting a super fast chip, and in June they still do not know what aniso method they want to use ? A bit confusing, is it not ?

But fear not. In October 1999, there was - apparently - solid information supporting the idea that the Glaze 3D was real. The article states quite clearly that Bitboys have already earned themselves the title of "King of Vaporware", so that might be bad news for those who actually wanted to believe the hype. That nonwithstanding, we learn interesting numbers like 15 MPixels/s, 4 pixel pipelines, 15 million triangles per second, and more.

But it does not stop there, oy no. On January 24, 2000 (uh, a year and a half later then, right ?), we learn that the Glaze3d now includes XBA - for Xtreme Bandwidth Architecture. The press release touts this new development as a landmark in the Glaze3D project, which hasn't really gotten anywhere near store shelves yet, it must be said.

In any case, more numbers have been thrown at the public : bandwidth of 12Gb/s (20Gb/s in dual chip - while todays' FX 5900 has just reached 27Gb/s), 9MB of external RAM and "very high resolutions" (hey, are we still compatible with VGA here ?), and Quake III at two to three hundred fps. Gosh, my eyes are watering with delight.

To put things in perspective, in the year 2000 nVidia had the GeForce 2 out, with a 5.3Mb/s bandwidth. ATI was puttering along uselessly and 3dfx was getting mired in vaporware over the Voodoo 4. It was a moment of great opportunity, and Bitboys could have taken a huge share of the market - had they actually produced something worth the specs they so loudly touted.

That was not to be, but meanwhile the company (which still hasn't, as of this writing, produced any working silicon in ANY stage) declares having received over 5 million euros in investment money. Something tells me that 5 million euros for a patent on a wild idea is not something the investors will content themselves with.

Meanwhile, May 15th 2001, a well-written editorial tries to hype the idea that 3D gaming is better for the PC, whatever the console guys would have us believe. Great news guys, how about actually taking part in the business ?

In September, another editorial underlines the critical importance of bandwidth in 3D. The article is right on the nail, and is quite right generally speaking. But there still is no Bitboys silicon anywhere.

So, to resume the situation, in 1999 Bitboys already had a history of ridicule. The press release for the never-to-be-seen Glaze3D was on their site for no less than two and half years, without any silicon to sell. During that time, nVidia trashed 3dfx with its GeForce 2, brought to market the GeForce 3 and, in February 2002, launched the GeForce 4 with a bandwidth of 10Mb/s (right, it's less than 20, but at least it exists).

Now, since a few million euros have been invested in this jokers' shop, they have probably felt a bit of pressure to actually sell something other than press releases. And they just might have done it, if this article has any more meaning than the previous ones.

In the end though, this company has been even less than a Rambus without the lawsuit department (at least Rambus has their product in every Playstation on Earth). Their history, for the moment, is having taken nVidia's specs for every major new GPU and publishing a spec sheet with better numbers. Yet, since 1998, they have not sent out a single chip for review. Now, it would seem they have rearranged their target market to something that might be more feasible : the mobile market.

With this new entity being lauded by Bitboy's marketing crew (which has remarkably toned down its usually over-the-top enthusiastic prose to stick to more technical and feasible data, albeit still slightly enthusiastic) is DX8-compatible and promises 9x anti-aliasing at resolutions of 4000 x 4000 pixels (ok, that is no longer VGA-compatible for sure). So, basically they admit they cannot yet make a DX9 GPU, but they still can't help making a press release without hyping something outlandishly. Old habits die hard.

In the end, despite the proposed reduction in design time "using already existing, efficient designs" (like nVidia's NForce ?), Bitboy's had better hurry something to the market before nVidia gets out its next Go product. Indeed, there are rumors that the next mainboard chipset will include an on-board FX5600 chip. Since the NV31 is DX9-compatible, the Axe will be put to the technological bin by default.

How ironic. Five years of hype and rabid imagination, and it all ends in the waste of an already-outdated product.