The OT spec defines encoding ID 0 for the Unicode Platform (PID=0) as "Unicode 1.0 semantics". If I'm writing a 'name' table with an entry for PID=0 / EID=0, how would I encode (or, more to the point, restrict) the characters??
I've found this "reconstructed" UnicodeData file that was reverse-engineered by Ken Whistler in 2004, but that's all I have to go on ...
Would simply dis-allowing characters that are not in Ken's reconstruction be sufficient to abide by "Unicode 1.0 semantics"? ... or is there something more that I need to do??
If all that is called for is to disallow characters, then ... what is the difference between Encoding ID's 4 and 6 for Platform 0?
* EID=4 is "Unicode 2.0 and onwards semantics, Unicode full repertoire ('cmap' subtable formats 0, 4, 6, 10, 12)" and
* EID=6 is "Unicode full repertoire ('cmap' subtable formats 0, 4, 6, 10, 12, 13)"
Aside from allowing cmap subtable 13 for EID=6, is there any difference between these two? Do I need to (groan) plow through an ancient Unicode standard?
BTW, I've found significant number of Open Source fonts that still use EID 0 and 4 for the Unicode platform. However, they pale in comparison with cmap tables for the Windows platform. Windows/Unicode BMP (PID=3, EID=1) is by far the most prevalent (96.7% of the fonts have that for the 'cmap' table), while only 16.7% carry a Windows/Unicode Full 'cmap' table.