# UTF-4088 encoder and decoder Status: experimental V1 implementation contract. ## Identifier namespaces The 16,606 front-end records retain identifiers `0x110000..0x1140DD`. V1 packed input-state identifiers use the separate interval beginning at `0x200000`. This prevents collisions between the front-end registry and encoded input states. ## Input packing A 21-bit packed input contains x (3 bits), y (4 bits), pressure-minus-one (6 bits), and normalized voltage (8 bits): | Bits | Field | Valid values | |---|---|---| | 0–2 | x | 0–7 | | 3–6 | y | 0–11 | | 7–12 | pressure minus one | 0–49 | | 13–20 | normalized voltage | 0–255 | Bits 21–31 are reserved and must be zero. Values are logical normalized application inputs, not raw electrical readings or physical pressure. The identifier is `0x200000 + packed_input`. Decode subtracts the base and validates every field. Invalid coordinates, pressure values, reserved bits, front-end IDs, and out-of-range identifiers are rejected. ## API Include `encoder_decoder.hpp` and use `pack_input`, `unpack_input`, `encode_input`, and `decode_input`. `symbol_from_input` accepts a packed V1 input word and returns the disjoint experimental identifier. ## Limits This is an identifier mapping, not a Unicode encoding and not a 4,088-bit integer. It does not encode semantic meaning. Semantic annotations remain separate registry data with provenance and review status.