> I wanted it to be stateless and functional, whereas with space-partitioning trees, don't you need to maintain the tree (and not generate it on the fly for it

> I wanted it to be stateless and functional, whereas with space-partitioning trees, don't you need to maintain the tree (and not generate it on the fly for it to be scalable)? Yes, but it's essential to maintain some sort of summary or index data structure to make the method scalable. A linear scan may be stateless, but that's hardly an issue when the method won't scale beyond a few hundred examples. > As for your second point, could you elaborate on why the movie example is non-metric? No triangle inequality.