From kenh@microsoft.com Mon Jun 22 11:09:25 1998 Received: from burdell.cc.gatech.edu (root@burdell.cc.gatech.edu [130.207.3.207]) by lennon.cc.gatech.edu (8.8.4/8.6.9) with ESMTP id LAA17264 for ; Mon, 22 Jun 1998 11:09:23 -0400 (EDT) Received: from wheaten.hitl.washington.edu (lweQKGe2hyjL2Ov9YQ4dZm7sE+/WoEIn@[128.95.73.60]) by burdell.cc.gatech.edu (8.8.4/8.6.9) with ESMTP id LAA24378 for ; Mon, 22 Jun 1998 11:09:22 -0400 (EDT) Received: from mail2-b.microsoft.com (mail2-b.microsoft.com [131.107.3.124]) by wheaten.hitl.washington.edu (8.8.8/8.6.12) with ESMTP id IAA13073 for <3d-ui@hitl.washington.edu>; Mon, 22 Jun 1998 08:09:11 -0700 (PDT) Received: by mail2-b.microsoft.com with Internet Mail Service (5.5.2166.0) id ; Mon, 22 Jun 1998 08:08:31 -0700 Message-ID: <5F68209F7E4BD111A5F500805FFE35B9057977DF@red-msg-54.dns.microsoft.com> From: Ken Hinckley To: "3DUI (E-mail)" <3d-ui@hitl.washington.edu> Subject: RE: input devices Date: Mon, 22 Jun 1998 08:08:28 -0700 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2166.0) Content-Type: text/plain; charset="ISO-8859-1" Status: RO Drew wrote: A question to be asked, here, is: can a hand input device be created which provides sufficient accuracy for recording fine motions. [...] To summarize: Props make good input devices when matched to the task at hand, and are relatively easy to use by software. However, they do not provide a general solution to user input (what if I want to manipulate objects of very different sizes, but don't want to buy a tracker for each?). Hand input devices (eg. gloves) provide a model of the user's hand, which if accurate enough, I believe could be the general solution. My point: gloves may suck now, but I don't think they will suck in 5-10 years. I think we'll have to agree to disagree on this point :). Seriously, the fundamental problem with gloves is that, even if the sensing technology were perfectly accurate, there is no haptic feedback to guide your motions. So the reasons that I find gloves problematic for many position/orientation manipulation tasks will still be true in 5-10 years. Now, as Jeff has pointed out, there are situations where this doesn't matter or is a secondary issue (with the head-crusher being my favorite example). There are also situations where gloves directly capture a gesture of interest (such as pointing) or where the extra DF's of the joint angles may be useful. The issue of generality is a separate one -- a "3d ball" that lets you position and orient objects is also a general input device. (Having a physical device vs. having a "prop" that closely matches a particular virtual object is another separate issue). > The other key issue, which I don't think anyone mentioned yet, is that with > a glove you absolutely need a clutching mechanism if you want to orient > objects in any direction. With a hand-held tracker or prop, then a clutch > isn't necessary (but may be desirable depending on the precise task > constraints) -- so as a designer you have another option that I feel you > don't have with gloves. I don't understand what is so hard about clutching using a hand input device. Could I not simply press my fingers together to grab, rotate (move fingers and thumb in opposite directions, or twist wrist), open my fingers to release, grab, etc? Of course, the software should support this technique, and not change the axis/origin of rotation while it perceives a single object rotation actions. I can think of a lot of physical objects that require clutching (most bigger than my hand). Also, I cannot rotate a small physical object smoothly over one rotation in any direction without pausing, unless I use two hands. Could I not do this with two, accurate hand input devices? What am I missing here? All I was trying to suggest is that you can achieve any orientation by tumbling an object in your fingers, but the same is not true if you try to move a single hand to any orientation. I'm not suggesting that clutching in itself is a necessarily good or bad thing, or inherently difficult. My point is that it's often assumed that clutching is *necessary* for 3D manipulation, but this does not have to be the case. For the specific example of my doll's head interface, there is no clutching button for the left-hand object (the doll's head) as we found that it caused problems and had very little benefit. My forthcoming article in ACM Trans. on CHI will include a discussion of this issue. Ken Ken Hinckley Microsoft Research One Microsoft Way Redmond, WA 98052 (425) 703-9065 kenh@microsoft.com