Helen Borrie, The Firebird Book, 2004, USA, page 175:

ISO8859_1 (LATIN_1) and WIN1252
The Firebird ISO8859_1 character set is often specified to support European languages. ISO8859_1, also known as LATIN_1, is a proper subset of WIN1252: Microsoft added characters in positions that ISO specially defines as NOT A CHARACTER (not "undefined," but specially "not a character"). Firebird supports both WIN1252 and ISO8859_1. You can always transliterate ISO8859_1 to WIN1252, but transliterating WIN1252 to ISO8859_1 can result in errors.

Carlos Henrique Cantu, Firebird 2, 2006, Brazil, page 125
The ISO collations order by "dictionary sort technique" where the space and other accent characters are treated in a different way what results in a "strange ordering."

Collations by Peter Gulutzan
http://www.dbazine.com/db2/db2-disarticles/gulutzan1


1. Are those collateral effects of the ISO8859_1 character set (and PT_BR collate) ordering real? And if they are, do they still persist on Firebird 2?

2. Is there any problem in using the WIN1252 character set (and WIN_PTBR collate) on a Debian Linux Server? The clients will be Linux with Java and C++ desktop applications and Windows XP Pro with Java and Delphi 7 desktop applications.

Best regards,

Valdir


[Non-text portions of this message have been removed]