Float16

Logic for converting half-precision (16-bit) floats to and from bytes.

Source: https://stackoverflow.com/questions/6162651 (The author says it's public domain.)

The documented author's "extensions" have been removed for clean round-tripping to and from Float.