Investigate if there's any sane reason to implement more optimal Simple-family codec, such as using dynamic-programming to find optimal block-length, such as in "Optimal Packing in Simple Family Codecs" paper. The other option would be look at powturbo's VSimple if there's any useful ideas that could be implemented here (some of the ideas in the TurboPFor use methods that are not available/useful in Java).
Investigate if there's any sane reason to implement more optimal Simple-family codec, such as using dynamic-programming to find optimal block-length, such as in "Optimal Packing in Simple Family Codecs" paper. The other option would be look at powturbo's VSimple if there's any useful ideas that could be implemented here (some of the ideas in the TurboPFor use methods that are not available/useful in Java).