#78 Refactor character set encoding/decoding

[Andrei Kapustin]

Currently, it's a CharacterSet that does the encoding/decoding between characters and byte sequences. This model doesn't allow encodings that use shift-states, etc. What needs to be done is 1) defining an Encoder/Decoder interfaces (nested in CharacterSet class), 2) adding factory CharacterSet methods to create new encoders/decoders and 3) moving the actual chars <-> bytes conversion logic into these en/de-coders.