Meeting Date: 17.05.1999
What I did last week: I gave up once again on CoRgi. So I thought a lot about 2D selection and made some pretty drawings and came up with all these ideas about how to disect an arbitrary shape. In principle I decided to deal with line-segments that make up a closed curve. I wanted to allow intersecting curves, so that complicated things quite considerably. When the shape is purely concave there is no problem, as one can adopt a bisection-approach in 2D, but I cannot assume such simplifications. So then after a lot of thinking I woke up at 6:30 in the morning (pretty amazing as such) and suddenly had this idea of counting the numbers of crossings with the outter contour of the shape: Any point inside crosses the boundary an odd number of times, any point outside crosses an even number of times (if at all, but 0 is also even). Details are as always on the worksofar pages.
What I'll do this week: I will verify my ideas by coding the whole thing. I also had some ideas about increasing efficiency, because we wouldn't want to test all vertices in a world for selection, but only the ones in the vicinity of the selection-spline. I think I found a way of calculating a circle that encloses all vertices in a spline and has a nicely tight fit. I assume that most selectionsplines will have some kind of circular shape, so that it will make a good approximation.