[spec] Clarify when to use which escape mechanism

[brawer]

In the Feature File Syntax specification, the current wording of section 9.e is a bit unclear about which escape mechanism (single-byte or double-byte) is applicable to what encoding. The current spec ties double-byte to “Windows” encodings. But in OpenType, the Windows platform can have encodings other than UTF16-BE. For example, platformID=3 + platformEncodingID=5 is EUC-KR/Wansung, which is a variable-width encoding so double-byte sequences won’t actually work.

As of 2017, this doesn’t matter much anymore, but perhaps the spec should say “UTF-16” instead of “Windows”.

[kenlunde]

Those legacy Windows encodings are still byte-based, like the Macintosh ones, and many of them are actually the same across platforms. As such, they would still use the \hh escaping mechanism, not the \hhhh one, which is for Unicode 'name' table strings. I agree that the wording can be better and more tightened.

[readroberts]

Agreed