3 ms·
can anyone clarify "There is no guarantee that removing a point increases the area of the adjacent triangles; thus, the algorithm considers the effective area a
by andrewcooke 14y ago
can anyone clarify "There is no guarantee that removing a point increases the area of the adjacent triangles; thus, the algorithm considers the effective area as the larger of the associated triangle and the previously-removed triangle" - it does't appear in the original paper (afaict) which seems to emphasise that the method is point-based (implying no such "memory"). and i can't see why it's needed.
it's clearly there in the source, but i don't understand the comment there, either. what am i missing?! (i understand the lack of a guarantee of increasing area - what i don't understand why the algorithm should "care"; once the point is gone, it is gone...).
- mbostock 14y agoThe line from the original paper is in section 3.3: "If its calculated area is less than that of the last point to be eliminated, use the latter's area instead. (This ensures that the current point cannot be eliminated without eliminating previously eliminated points.)"
- andrewcooke 14y agoah, thanks. and i think i understand now - it doesn't affect filtering, but it makes the precalculated list have a monotonic value that can be used to select a "degree of smoothing". maybe.