From: kirlik@chmsr.gatech.edu (Alex Kirlik)
Subject: Re: Can displays be "too real"?
Date: 21 Oct 90 07:47:17 GMT
Organization: Center for Human-Machine Systems Research


In article <9638@milton.u.washington.edu> dmark@acsu.buffalo.edu (David Mark)
 writes:
> some things deleted
>
>So to rephrase my original question, has the Virtual Reality community
>addressed the issue of how real the presentation should be, and when the
>presentation to the user should be simplified and selective?
>

I am happy to see someone raise this issue.  I think it's important.

No, I don't believe they have.  While I am pleased to see advances in VR
along the lines of realism, I have often wondered what is really driving
this work.  One thing that's absolutely amazing about perception is how
little realism you really need to "see" something; like the Grand Canyon
on a little piece of cardboard while sitting in your living room.  

I know this is missing the point of the original posting, which seemed to
be that abstraction (rather than pure realism) is sometimes better for
certain tasks.  I agree here too.

It strikes me that VR is predicated on achieving the phenomenological aspects
of perception, the quality of "presence," which may in fact be a very noble
and interesting task.  Yet, nowhere have I ever found any empirical evidence
to suggest that the phenomenonological side of perception is integral to the
cognitive side, i.e., to selecting the appropriate behaviors.

I'm not contending that the feeling of "presence" is not integral to behavior,
I am just saying that there is no firm evidence to show that a profound 
feeling of presence is truly necessary to get tasks performed.

Perhaps my attitude reflects (too much) my own interests, which is improving
interfaces and human-machine interaction in real world systems.  If someone
would show me that pilots (say) perform much better when they report a
feeling of presence from their displays, than when they are aware that their
displays are just that, displays, then I would be the first one calling for
VR type systems in the cockpit, etc.

The fundamental question (when designing for performance, rather than for
entertainment or other purposes), hasn't really changed, it is still, "what
information is needed for best performance and in what format should it be
presented?"  I am happy to see the advances in VR, should we ever find that
the answer to the above question is that we need not only the information but
the phenomenonological aspects of presence too.  That may happen, but then
again it may not.  If I am correct that this is still an open question, then
to me VR remains a (very interesting) technological opportunity, but by no
means the solution to the many interface design problems we have today.

The problem, of course, is that it is not clear that presence for presence
sake is really such a good thing.  My guess is that presence will cut both
ways, that when things are going well they will go quite well, but errors 
and misunderstandings will be amplified in intensity as well.  So, our pilot
has presence for example:  Now, instead of thinking (incorrectly) that his
left engine is out, now he is "fully experiencing" (incorrectly) that his
left engine is out.

I guess the bottom line for me is that I like to see problem-driven rather
than technology-driven solutions to real world design problems.  But I realize
that the potential applications of VR are much broader than my own area of
interest.  I would, though, like to see comments on what is "really" driving
the move to realism and presence.  

Alex 

UUCP:   kirlik@chmsr.UUCP
        {backbones}!gatech!chmsr!kirlik
INTERNET:       kirlik@chmsr.gatech.edu