
* Integration to Slate
* Proper hash for strings
* Test normalization with the test suite provided by Unicode consortium
* Full case mappings
* In normalization:
  - isNormalizedTo: quick-checks
  - recomposeForForm: which trusts it's input being in NF(K)D,
    and not testing for isHangulLV, composing all three hangul in one pass
  - possibly NFC/NFKC normalization which decomposes only those that don't
    compose back
  - FCD 'normalization' (for faster collation)"
* proper lexicographicallyCompare (should be renamed collateWith:)
* Add more/refine Unicode subsets
* Add CharacterCollection and some other things from (old) string.slate
* Bidirectional rendering (probably after there is UI)
* Enhance standard-compliance (x5)
* Enhance efficiency
