Exact-width lookup numerical systems

This is the public entry point for lookup-table numerical formats. It exports the generic codebook representation, the supplied catalog, proved catalog arithmetic, the configured ExecFloat.Codebook carrier, numerical automation, and #float_info support.

Import narrower submodules when defining another codebook family internally.

Unlike standardized interchange formats, a codebook is defined by its complete denotation table. Consequently, the table itself is the authoritative format specification; FloatLib does not silently infer arithmetic from bit width or element type.

Reference

Codebook.denote specifies the meaning of every stored word, including exceptional entries. nearestCode supplies nearest-entry quantization; catalog arithmetic is defined separately.