From: John Wann <john1@HMS01.HMS.UQ.OZ.AU>
Subject: Re: SCI: Visual Cues to Depth Perception

 joachim lous <joachim@KIH.NO> wrote:
>..... I was able to judge the distance to an object pretty well, It
> just din't feel REAL, because all the little cues we observe in real 
> life, and hardly know about, much less register consciously, 
> were missing.  
>

There is a difference in our perspectives that radically alter the 
weight you place on various cues to depth. If you are interested 
in trying to re-create "reality", then sure you can try to bundle in 
every feature of shading, haze etc.. that your graphics generator 
can handle, but ultimately you pay the cost in terms of real-time 
performance (another requirement of "reality"). - You simply 
have a classic speed-accuracy trade-off.  Personally I think the 
quest for "reality" is a fools errand, much loved by the media but 
of dubious applied benefit. 

Why do people want "reality" ?  What is necessary are robust and 
easily usable interfaces that transmit information and allow 
interaction. - This does not require any concept of "reality", 
merely the careful consideration of what features are essential to 
support human control in the task to be undertaken. The early 
days of AI were bugged by aspirations of building truely 
intelligent and creative machines, before the science came to 
ground and decided that systems that made smart judgements 
and had behaviour that was robust with a limited range was a 
sufficient goal.

Rather than throwing everything into the pot and producing a 
turbid stew, we favour an approach of laying out recipes of what 
needs to be included, and what can be left out, given the aims of a 
specific menu.

John Wann
Edinburgh Virtual Recipe Project