From: Maarten van Dantzich [maartenv@microsoft.com] Sent: Wednesday, November 29, 2000 5:40 PM To: 'bowman@vt.edu'; 3dui List Cc: Rex Hartson Subject: RE: notation for 3D interaction techniques Doug Bowman: > There are some real subtleties in the implementation of some of > these techniques that would be useful to developers if they were > available. >From this, I assume the intended audience would be developers re-implementing an interaction technique from a description in the literature? I would argue that it is not only important to describe how the interaction technique is best implemented, but also why certain details are present. In other words: what problematic behavior did you avoid with some of the subtleties of the implementation? I'd agree with Roy Ruddle, and extrapolate from his comments: start by building a collection of interaction technique specs, and then distill from it "Patterns of Interaction" [cf. Design Patterns], written up in a common style, with common elements like pseudocode, state transition diagrams, descriptions of gotchas, minimal qualities of input devices, etc. Whether or not it's possible to agree on a pseudocode notation is a big question in my mind. This would depend a lot on the commonalities of the different toolkits out there, and whether they have compatible models of input events, camera, scene graph, animation, etc. Would you really ever fit one VHLL on top of each of them? Discussion on what's desirable and/or achievable sounds like it'd be a great topic for a workshop, perhaps attached to I3D this Spring? >> Maarten.