Eszett

[Michael Everson]

On 12 Jul 2009, at 00:58, John C Klensin wrote:

>>> No capital letter in a multiple-case script is PVALID. Not
>>> U+1E9E   and not "A" (U+0041), "Å" (U+00C5), "Β" (U+0392),
>>> etc.
>>
>> Ah. Then PVALID means "the lower-case letters" and their
>> upper-cases are trivially "mapped" (linked) to them.
>>
>> Yes?
>
> More or less.  See below.

All righty.

>>>> I would make ss and SS and ß and ẞ equivalent.
>>>
>>> If we support (whether as a recommendation or a requirement)
>>> lower-  casing or case mapping of any other flavor,
>>
>> "Other flavour"? Please use plain English. I have no idea what
>> you mean by this idiom.
>
> Apologies for the shorthand.  One of the discussions that seems
> to be continuing is whether the case-related mapping activity
> should use the Unicode functions toLowerCase or toCaseFold.
> Since the latter, if applied to U+00DF, would transform it into
> "ss", an exception would need to be made if it were chosen.

I should favour toLowerCase then.

> Some of us believe that is one of perhaps several reasons why
> toLowerCase (which requires no exceptions) should be used rather
> than getting involved with toCaseFold.

I agree with you there.

> Note that the real controversy at the moment is what the status
> of mapping operations should be under various circumstances --
> optional or required and, if required, when and whether there
> are exceptions.

I gathered that.

>>> I certainly hope that U+1E9E will be mapped into U+00DF.
>>
>> So do I, whether or not either is "mapped" (linked) to
>> "ss"/"SS"/"Ss"/"sS".
>
> The nature of the beast is that, if mapping occurs at all,
> either:
>   U+00DF is PVALID
>   U+1E9E maps to U+00DF
>   "s" is PVALID
>   "S" maps to "s"
> or
>   U+00DF maps to "ss"
>   U+1E9E maps to "ss" or is DISALLOWED entirely
>   "s" is PVALID
>   "S" maps to "s"
> If there are other possibilities or combinations, I cannot
> imagine them.

I very much FAVOUR the first option, and I very much OPPOSE the second  
option.

>>> It appears from my reading of the Mapping draft and
>>> UnicodeData.txt   that mapping (and the mapping from "S" to
>>> "s" will happen without   making any special provisions, so
>>> we should be ok.
>>
>> I'd like to have this confirmed.
>
> If it is not correct, I think we would all like to know that.

...

Michael Everson * http://www.evertype.com/