The tiny floating-point encoding functions we've been using are very useful. Please add them to the library.