CompuServe Thread

ICP Ideas

2 messages in this thread
#41695From: SyndesisJul 20, 1992 10:48 PM
1. A good idea, I'll drop of the characters. I've been trying to think of ways to address these limitations and incompatibilities. 2. Again, same question. I've pressed the creators of LightWave and Modeler to make the name requesters the same length. The old InterChange had a fixed list of about 64 colors, which most people regarded as a limitation. The new ICP has a stupendous number of color names, generated by an algorithm, not a table of pre-made values. 3. Yes, it's in the manual. 4. Yes, this would be a good feature. 5. This would require artificial intelligence, which means that no matter what we'd do, it wouldn't work right all the time. 6. Yes, it would be easier to remove exact duplicate polygons. 7. There's no absolute scale for object size. No one program says "1.0" means "1 meter" although LightWave gets close. 8. LightWave requires a single point in an object, even if it's not part of any polygon in the object. Since the Imagine sub-object had no points then the LightWave object is given a single point and no polygons. 9. One man's extra point is another man's work-around for a bug in a rendering program. Maybe, someday. 10. Our InterFont product uses spline curves to make 3D text, and it uses an algorithm similar to this. However, in a polygon-based object, how could you regenerate the spline-knot information? 10b. Yes, this is exactly how InterChange Plus's Surface files work, except for the part about "vary this surface by 10 percent."
#41774From: Erik FlomJul 22, 1992 12:25 PM
Thanks for your responses to my suggestions. As I had guessed, some are already implemented in your software – I WILL have to buy Interchange Plus someday :^). The incompatibilities amongst the 3-D package interfaces IS vexing for both user AND developer. I don't blame you guys for making names longer than the field it fits in, and actually find the extra information useful (somtimes). When I sent my list of comments, I realized that some things I wanted implemented were not possible (at least easily), so thanks for at least looking at them. As for the use of a spline algorithm to add points to an object: You probably could just use defaults for the spline-knot info. Though the curve might not fit perfectly, it would be a decent simulation. The user could then touch up the curve in a modeling program. A case in point was the Fender Stratocaster object – I just about doubled the number of points along the curved outlines. This was an involved process of creating the points, then adding them to the polygon, then double-checking to make sure the curve was correct. It would have been nice to automate the first 2 steps – which I had to repeat some 50-80 times. I like the fact you have Surface files for Interchange – it would be useful if it was tolerant of variances in the input surface values, then the user wouldn't have to worry if he/she had tweaked the surface since the library was created. Though it might already work that way. Thanks Again, Erik Flom