Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!samsung!uunet!ogicse!milton!wex@dali.pws.bull.com From: wex@dali.pws.bull.com (Buckaroo Banzai) Newsgroups: sci.virtual-worlds Subject: Semantic space, navigation Message-ID: Date: 21 Aug 90 16:46:34 GMT References: <9007250107.AA01311@hitl.vrnet.washington.edu> Sender: hlab@milton.u.washington.edu Organization: Bull Worldwide Information Systems Inc. Lines: 79 Approved: hitl@hardy.u.washington.edu In article brucec%phoebus.phoebus.labs.tek.com@RELAY.CS.NET (Bruce Cohen) writes: In the large-scale kinds of cyberspace people keep talking about, I think the only solution is intelligent "guide" programs, which can build up a navigation space with a suitable topology to match what a user needs at any given time. That's an interesting idea - a specialization of the general "assistant" type programs. I hadn't thought too much about that. I have been concentrating on the kinds of artifacts we might put into cyberspaces to aid navigation. We tend to think of an overview like a map, which shows all the important spatial relationships down to some minimum scale. This doesn't work well for high dimensional continuous spaces or highly connected non-continuous spaces. Overviews in high dimensions tend to be more like document outlines than maps. I agree an overview is not a panacea, but I do think one wants to have some kind of map-like view which allows one to put local detail into global context. This is why I'm interested in generalized fisheye views, which have this as their explicit purpose. The user still needs to have some sort of mental model of where she's going to go when she teleports, or steps through a trapdoor to the next level down. People can accept nonlinear movement, but they still need a map of some sort that they can internalize at some level of detail. Yes and no. Teleportation is used for those time when the user says "I don't care where I am or where my goal is, but I know my goal is X - take me there. No real map is needed for this; just a unique method of identifying targets. On the other hand, you're right that people cannot navigate to any significant degree unless they have *some* internal map of where they're going. The key point of all this discussion, which I doubt very much we disagree on, is that it's very important to consider the cognitive and sensory capability of the person inhabiting the virtual world when designing it. Developing a taxonomy of design or implementation spaces for these worlds is a useful project, but we can't let that be the end of our thinking about the use and variety of space in Virtual reality. The real-world 8-) virtual reality designer is going to have to deal with (at least) two kinds of space, the world design space and the user navigation space, and the mapping(s) between them. Very true. Here we completely agree. As I've mentioned before, the idea behind semantic spaces is to develop a fully-general, extremely powerful theory with which people can unify information about cognitive and technological limits to design the cyberspaces they want. [...] the landscape you're trying to navigate through is probably changing on several different time scales; some of those changes are local to your current locality and on a time scale at least as fast as or faster than your own movements. The map keeps changing, inpart due to your own actions, but only in part. This turns out to be incredibly disorienting to people. Our experience (admittedly *very* small) is that people can deal with themselves moving, or the landscape moving, but don't do well with both, unless the landscape movement is well-constrained. You can see this in two simple systems. In a system which allows users to get their viewpoint up to relativistic speeds, people tend to lose track of what they're doing when the terrain they're passing over starts changing. Also, try one of those games where you have to land a plane on a moving target (like an aircraft carrier) - very frustrating. I agree. We can learn an awful lot from the way such annotations are handled in thematic maps which try to merge several different kinds of information. I woudl like to see more people in the hypertext/vr area studying the way physical and thematic maps are designed; there's a lot of useful prior art there. I'm not sure what you mean by "thematic maps." Can you give some examples and a reference or two? -- --Alan Wexelblat phone: (508)294-7485 Bull Worldwide Information Systems internet: wex@pws.bull.com "Politics is Comedy plus Pretense."