spec text guidance for implementors

[phoddie]

Implementing float16 support is somewhat challenging because compiler support for 16-bit IEEE floating point is not yet universal. Moddable encountered this building for certain embedded targets and Windows. We solved this for XS by following Mozilla's example of adapting code from the excellent Half-precision floating-point library library.

In some cases ECMA-262 provides notes to implementors, particularly for floating point operations. These are some examples:

• "Note 2" in 6.1.6.1.6 Number::remainder
• "Note 3" in 6.1.6.1.20 Number::toString
• "Note" in 21.3.2 Function Properties of the Math Object

Given the uneven status of native compiler 16-bit IEEE floating point support, it would be helpful to include such a note in the spec text of this proposal.

(This is somewhat related to #12)