Re: [idn] Namprep-02: ß.com

[Harald Alvestrand <Harald@Alvestrand.no>]

At 10:09 02/02/2001 -0800, Rick H Wesson wrote:

>Nameprep ß.com = ss.com
>
>It's illegal to register ss.com as a ml domains because its only
>plain ascii after nameprep.

"this" can mean 2 things, depending on how I interpret what I see.....

Either this is Unicode x00DF:
00DF;LATIN SMALL LETTER SHARP S;Ll;0;L;;;;;N;;German;;;
Or this is Unicode x032B:
03B2;GREEK SMALL LETTER BETA;Ll;0;L;;;;;N;;;0392;;0392

Both of these are lower case letters (Ll).
None have decompositions, canonical or otherwise.

The relevant text from nameprep-02 that produced the "sharp is ss" is:

>3.2 Additional folding mappings
>
>There are some characters that do not have mappings in [UTR21] but still
>need processing. These characters include a few Greek characters and
>many symbols that contain Latin characters. The list of characters to
>add to the mapping table were determined by the following algorithm:
>
>b = Normalize(Fold(a));
>c = Normalize(Fold(b));
>if c is not the same as b, add a mapping for "a to c".
>
>Because Normalize(Fold(c)) always equals c, the table is stable from
>that point on.

(Aside: I don't understand how this produced the entry "00DF; 0073 0073; 
Case map" in the tables, since there is no string that folds to 0x00DF.
It's "logical" in some sense, since 00DF mapped to UPPERCASE is 0073 0073
according to the Special Casing datafile of Unicode. But "Fold" does not 
map to uppercase, it maps to lowercase. And then 00DF maps onto itself.
Specification seems incomplete.)


>How can anyone register  ß.com or will it be unavailable?

According to the current nameprep doc, it cannot be entered into a 
zonefile, so it will be "mechanically unavailable".

The Greek Beta will be available.


--
Harald Tveit Alvestrand, alvestrand@cisco.com
+47 41 44 29 94
Personal email: Harald@Alvestrand.no