- If null-radix? exists it might as well be used in radix-append.
- Gen-leaf-fold is not a leaf fold, it doesn't use 'leaf' at all. It folds an inlined iota. Since 'proc' will inevitably use 'pos' to extract a leaf item, and since the purpose of the fold is to factor out common work, this extraction should be in the fold and the item passed to 'proc'.
- Similarly, the level-1 fold also folds an inlined iota. It should extract the level-1 vector from the radix and pass it to 'proc-2'. It should also remove its current return asymmetry by ditching 'proc-1' and the first branch entirely.
- The lambda of leaf-for-each should return 'acc' to suppress the return of 'proc'.
- The 'begin' in the lambda of leaf-map is pointless.
- Once the level-1 fold stabilizes it should be used to drastically simplify radix->vector.