IDNAbis discussion style, mappings, and (incidentally) Eszett

Patrik Fältström patrik at frobbit.se
Mon Dec 3 01:25:45 CET 2007


On 30 nov 2007, at 08.28, Paul Hoffman wrote:

> I had the same perception that Erik did. If I had known up front  
> that there were only six exceptions, I probably would not have been  
> so disappointed reading the document. Maybe it would be good from a  
> marketing perspective to say up front the problem you had but, even  
> before that, say that there are only six exceptions.

The layout of the rules in the tables document is as follows:

A. Look at what GeneralCategory the codepoint is (list good ones)

J. What GeneralCategory need context dependent rules? (extension of A,  
that might be ok)

B. Look at whether normalization of the codepoint is stable or not  
(only stable codepoints are ok)

C. Look at whether casefolding of the codepoint is stable or not (only  
stable codepoints are ok)

D. Look at the property of the codepoint (ignore codepoints with  
special properties)

E. Look at the script of the codepoint (ignore codepoints of certain  
scripts)

F. Look at the block the codepoint belong to (ignore codepoints of  
certain blocks)

K. Unassigned codepoints


Then special cases:

G. Take care of ASCII LDH for historical reasons

H. Pick exceptions (6 of them)

I. Take care of HAN <--------- Bug in document!!! This is Y


Then calculate whether the codepoint is ALWAYS, MAYBE, NEVER, CONTEXT  
or UNASSIGNED.

        Patrik



More information about the Idna-update mailing list