Bill, there is some relevence to what you say about reformatting the data, but at some point, in just about every 3d pipeline I've ever messed with (hardware or software) the data is reduced to a set of polyons and the various material data etc. anyway, so I really don't see it as a problem. In OpenGL, if (or while) the model remains static, you can build display-lists and just render the lists (which are stored on the video hardware) until/unless something changes. So when you're just rotating, translating or scaling the entire model (or moving the camera around, for example), the app simply re-uses the display list each frame. If the model is animating, it either passes in all the data each frame or rebuilds the display-lists each frame... either way should still be faster than software rendering. While there may be some overhead involved while animating, 90%+ of Poser users are just rendering still shots. And even when they are animating, the time you actually have the animation active is pretty small compared to the time you spend setting up poses, and rotating the view around to get clothing conformed, etc. In short, I don't see any disadvantages to offering hardware rendering in preview mode (aside from the time that the programmers need to initially implement it) and there would be HUGE advantages relative to making the application more useful/productive.

Cinema4D Plugins (Home of Riptide, Riptide Pro, Undertow, Morph Mill, KyamaSlide and I/Ogre plugins) Poser products Freelance Modelling, Poser Rigging, UV-mapping work for hire.