No one likes a muscle-head who can't carry on a decent conversation, and thus ATi wanted to make sure their beast wasn't only a powerful one, but also one able to handle the visual details well.
To achieve this, right away 32bpp-color support was mandated for the Aurora, as well as hardware support for bump-mapping, anisotropic filtering, and DirectX 7.0's implementation of S3's texture compression API.
By adding what are likely the three most popular visual quality features supported in games for late '99 and the first half of 2000 into the Aurora's native featureset, ATi guaranteed that they wouldn't have to rely on optimized software debuting on the market place later than the Aurora's market lifetime would equal.
Let's face it, no matter if you're a GeForce256 owner or an ATi Aurora owner, your card will probably not be cutting edge in a year. But if you can buy a card that offers massive performance gains in today's software right out of the box, rather than having to wait for game developers to support your card's new features (as with the GeForce256's impressive hardware transformation and lighting) then you'll likely wind up using and enjoying the card more right away.
This was the mindset of ATi heading into the Aurora's development, nd it's eerily similar to 3dfx's current position on features such as transformation and lighting as well.
With their Napalm project, due in early 2000, 3dfx has explained that due to a lack of perceived support until the second half of 2000, transformation and lighting is a feature they deemed unnecessary to support in their next series of video accelerators.
Unfortunately, we tend to agree with this stance at the moment, as games that include support for T & L shipping this year seem to be few. Even The Whole Experience, NVIDIA's Dagoth Moor Zoological Gardens demo provider, has designed their game "Experience" to work equally well with the current crop of accelerators. It will be a while before the installed base of hardware T & L supporting cards is large enough for publishers to green light games that take full advantage of the feature.
There are dozens of ongoing game projects that support some degree of T & L of course, including the work being done to id software's Quake3: Arena, Shadowman (patch pending) and Messiah.
Getting back to ATi's original mandate for the Aurora project, we find that the last primary point of the accelerator's goals would be speed.
In NFL football games speed kills, and in the wars between video accelerator cards it's equally important. Therefore ATi made a fateful decision for the Aurora's design, one that made sense on a variety of levels throughout the company and the industry: