In discussing these issues among Biblical Hebrew implementers, content providers and users, I have had to explain repeatedly why UTC doesn't want to consider this. It is completely obvious to them that this is the right solution. Even on explaining the impact on normalization, the response is that there is no impact since implementations and content using Unicode do not yet exist.
Indeed, but the UTC doesn't want to change the normalization stuff even where there are obvious errors, for philosophic reasons, I suppose. I mean who are all the implementors who depend on these tables? Often Unicoders have claimed "existing implementations" even where none can be shown to exist. Now Ken tempts us with:
"This is just one more in the accumulating pile of little problems in the decompositions locked down by normalization that will eventually result in the committee going "Spaaannggg!" and agreeing to publish and maintain a separate, corrected list of equivalences "As She Oughta Been" which are not constrained by the formal stability guarantees of UAX #15 normalization forms."
I'd like to understand how deprecating a character and adding a duplicate one with the right properties differs from deprecating a version of UAX #15 in favour of an Oughta-Been table.
:-)
I think it would be better to create a new character for this purpose than to use ZWJ in yet another way.
I suppose CGJ is tempting. -- Michael Everson * * Everson Typography * * http://www.evertype.com
