From bowman@cc.gatech.edu Mon May 24 13:29:19 1999 Received: from burdell.cc.gatech.edu (root@burdell.cc.gatech.edu [130.207.3.207]) by lennon.cc.gatech.edu (8.9.1/8.9.1) with ESMTP id NAA20092 for ; Mon, 24 May 1999 13:29:16 -0400 (EDT) Received: from asbestos.hitl.washington.edu (hitl-new.hitl.washington.edu [128.95.73.60]) by burdell.cc.gatech.edu (8.9.1/8.9.1) with ESMTP id NAA23188; Mon, 24 May 1999 13:29:13 -0400 (EDT) Received: from burdell.cc.gatech.edu (root@burdell.cc.gatech.edu [130.207.3.207]) by asbestos.hitl.washington.edu (8.9.3/8.9.3) with ESMTP id KAA06406 for <3d-ui@hitl.washington.edu>; Mon, 24 May 1999 10:25:19 -0700 (PDT) Received: from lennon.cc.gatech.edu (bowman@lennon.cc.gatech.edu [130.207.9.20]) by burdell.cc.gatech.edu (8.9.1/8.9.1) with ESMTP id NAA22678 for <3d-ui@hitl.washington.edu>; Mon, 24 May 1999 13:25:17 -0400 (EDT) Received: (from bowman@localhost) by lennon.cc.gatech.edu (8.9.1/8.9.1) id NAA19789 for 3d-ui@hitl.washington.edu; Mon, 24 May 1999 13:25:16 -0400 (EDT) From: bowman@cc.gatech.edu (Doug Bowman) Message-Id: <199905241725.NAA19789@lennon.cc.gatech.edu> Subject: Re: Virtual vs. real manipulation To: 3d-ui@hitl.washington.edu (3D UI List) Date: Mon, 24 May 1999 13:25:16 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Status: RO Thanks to everyone for their insightful and interesting comments regarding my question as to the current performance differences between virtual interaction and its real world counterparts. I think it's instructive that very few people actually answered the question directly, instead choosing to point out that perhaps the question was flawed because matching real-world performance is not the goal of VE interaction research. This is also the way that I answered the question when it was posed to me. However, I also think that we ought to be careful with such responses, because comparing virtual and real performance *is* a useful benchmark, as several people pointed out. As far as I know, no one has actually published research comparing the two for object manipulation, although it has been done for navigation (comparing a subject's spatial knowledge of an environment when training was done in a real vs. a virtual space - see Rudy Darken's VRAIS 98 paper and Waller et al in Presence 7(2)). I think it would be interesting to put together a technologically optimal VE system with state-of-the-art graphics, tracking, and haptic feedback and compare this to the real world for something like the block stacking task, just to get a feel for where we are. That being said, I realize that very few of us are researching new technology per se. Most of us are working on "magic" interaction techniques that make up for the deficiencies of the technology we're working with. It would also be interesting to include some of these techniques in the evaluation, allowing a reasonable level of magic (the system is not allowed to know the final goal state but is allowed to help the user with constraints, snapping, etc.). I think it's possible that we might find that such techniques do the job faster than real-world interaction, depending on the task (task complexity would be another interesting dimension to explore). Finally, I want to quibble a bit with Matt's definition of real vs. magic VE interfaces. He distinguishes these based on whether the user is taking anything out to the real world (needs a reality- based interface) or whether the VE is an end in itself (can use a magic interface). I would argue that there is a third class in which the VE is used to create an end product but the low-level process of creating that product is not important to the user. This category of applications can also use magic techniques. An example is my gorilla habitat design system. The user is creating a new habitat design, which can be saved and used in the real world. However, the technique the user uses to move a tree to a new location is not an integral part of the product - he's only concerned with the tree's final location. Therefore, magic techniques like HOMER or Go-Go can be used for this task. Probably most of the categories of VE applications fit into the classes that can use magic techniques. However, most of the real-world usage examples of VEs involve some sort of training, and therefore need natural interaction. (over-generalization) Comments? --Doug -- Doug Bowman, Ph.D. Candidate College of Computing, GVU Center, Georgia Tech Room 388 CRB, (404) 894-5104 bowman@cc.gatech.edu http://www.cc.gatech.edu/~bowman/