If you're at UC Davis, Danyel, we can get together sometime to discuss this.

I disagree with your evaluations of 3D -- or rather, multidimensional --
presentation of information, having worked extensively within immersion spaces
where the virtues of novel interfaces become evident.  Is this for everyone?
Clearly not.  But it works for those fortunate to have the experience.

bob


Danyel Fisher wrote:
>
> Bob,
>
> I'd like to hit a few of these points. I've been spending a number of
> years looking forward to being convinced that some sort of 3-D space
> could work as an information space, and haven't found much yet.
>
> I'll admit that I haven't tried anything bigger then a large monitor,
> and could be missing part of the game there. On the other hand, I've
> seen perfectly servicable, quite useful 3D on standard 15" monitors, for
> a specific range of problems.
>
> I am fully convinced that 3D is great for scientific visualization and
> imaging. I want to rotate that aorta to the right angle so I can zoom in
> on the clog. I want to watch the blood flow past the obstruction,
> rendered as a vector field of floating bubbles.
>
> But I see this as being very different from information visualization.
> I've seen few 3D renderings to abstract things--data that doesn't have
> an obvious real-world analog--that looks even remotely good, or
> understandable.
>
> To summarize: if we mean by "interface" a gui-like entity, something
> that allows generalized input and output for a large set of tasks, I
> strongly feel that 3D is unfit.
>
> If we mean "information visualization", then I'm looking forward to a
> convincing demo of a working system that provides information in that
> extra dimension.
>
> If we mean "scientific visualization," well, it would be foolhardy to
> argue against an established tradition of working technology.
>
> > But they'll be different from the
> > mnemonic and semantic allegories that currently pass for 3D interfaces.
> > Like natural language interfaces, natural perception interfaces will
> > process information the way we do, using the natural world as its model.
>
> I just don't understand _what_ this sort of rhetoric means. I've seen it
> a lot, and haven't been convinced that there is anything behind it. We
> don't have natural language interfaces, because natural language is
> hard. I can't see that the natural world leads to abstract information
> easily--like, say, visualizations of the internet.
>
> > Finally, I note that Magic Earth, which provides a 3D interface for oil
> > exploration and development, was just sold to Halliburton for ~$100MM.
> > Someone must like 3D interfaces...!  (See http://www.magic-earth.com/.)
>
> Highly specialized, highly representational (they draw _stuff_ that is
> already 3D). Not to mention, of course, that what _sells_ and what is
> _useful_ are often two different things. I wonder how thorough the
> adoption of this into Hallburton will actually be.
>
> Danyel