Hyphen Restrictions

Eric Brunner-Williams ebw at abenaki.wabanaki.net
Wed Jan 5 15:39:04 CET 2011


I don't think of the ??-- requirement as an implementation artifact of 
some bootstring encoding over ASCII mechanism.

That is, I don't think ??-- is generated by some algorithm acting on a 
string space generated over a repertoire, but as the pre-existing 
prefix which any bootstring encoding over ASCII algorithm must afix 
its output to.

The ??-- prefix exists externally to every candidate bootstring 
encoding over ASCII mechanism, as a device to signal something, 
perhaps a change of algorithms, perhaps a change of character 
repertoires, perhaps a change of sumptuary code or sartiorial norms 
amongst the IAB, we don't actually know (or care).

It just sits there and tells us it could signal a change to our string 
processing. It doesn't care if our repertoire is variable length 
encoded, fixed length encoded. It doesn't care if our bootstring 
encoding over ASCII mechanism is throwing darts or punycode or any of 
its percursors or successors.

Therefore the characterization of 8-bit bytes is the one I think is 
correct, where "?" takes on arbitrary values in the LDH (but not H) 
set, and for the moment, the values of "?" are {n, x}, ordered as "xn".

YMMV,
Eric



More information about the Idna-update mailing list