Re: The letter Sharp S and the English language

[Georg Bauhaus <rm.dash-bauhaus@futureapps.de>]

On 25.03.13 16:23, Adam Beneschan wrote:
> On Saturday, March 23, 2013 2:22:26 PM UTC-7, Georg Bauhaus wrote:
>> In case you remember a heated discussions of what � is,
>> whether it is an S-Z ligature or an S-S, and how to (not)
>> downcase "ACCESS", more evidence comes from Ireland of 1759,
>> in the signature of Arthur Guinne�,
>>
>> http://home.arcor.de/bauhaus/Ada/GUINNESS.jpg
> 
> That pretty clearly looks like two separate letters to me, although the two s's are in different styles.  But it isn't a ligature.  I'm not sure what your point is since I don't remember the original thread very well.

I'm investigating how Unicode enabled Ada can help me "export"
street names to Switzerland. Thus,

  To_Upper ("Xyz-Stra�e");  -- String or Wide_String

What interests me is whether or not this might or might not work in the
future, i.e. with Ada 2012, in the light of recent developments of
ISO/IEC 10646:

First, you'd typically not be writing '�' in Switzerland and instead
replace every occurrence with "ss". That's for both lower case and
upper case. (And also when using small caps). So, To_Upper's definition
won't help.

But! Since Ada 2005 there are two new twists. In 2008, ISO/IEC 10646
has published an official upper case character for '�', U+1E9E. And in
2010, official spelling (read: government; "amtlich") requires U+1E9E
in geographical names. These include street names.

http://141.74.33.52/stagn/Portals/0/101125_TopR5.pdf

Currently, GNAT's implementation of
 Ada.Wide_Characters.Handling.To_Upper
gives Wide_Character'Val (223) for To_Upper ('�'), AFAICS.

Unicode's CaseFolding.txt, if applicable, has two lines pertaining
to the matter,

1E9E; F; 0073 0073; # LATIN CAPITAL LETTER SHARP S
1E9E; S; 00DF; # LATIN CAPITAL LETTER SHARP S

So I'm wondering if Simple Case Mapping might mean that

  To_Upper (Wide_Character'('�'))

should return Wide_Character'Val (16#1E9E#) in Ada 2012.


('�' will thus continue to cause problems originating in web based form
entry fields and elsewhere, I'm almost sure. Just one out of many
experiences: a major fruit company's customer invoices have consistently
shown what looks like junk HTML right after "Stra" in my address for years.)