Well, “sparse” here is not dense, So, lots of and lots of room for various properties and such. Little things mean a lot. So instead of rounding, you use everything because maybe when things are combined that little thing will suddenly become the most important thing.
-Basically working with uncompressed data but also looking at the data after compression to see what’s important then back to uncompressed again.
—
Good question. I think what’s subconscious has to become conscious somehow.
—