Trent,I understand, this enhancement is certainly not the most important for all SF community.
I really need a specific collation (arabic) on some columns, though the tables will be in a French collation. These fields will be in NCHAR for unicode use, because I need good indices on these fields (few millions of rows per each table)
for example, a Person table will store name CHAR (30) in french script and arabic_name NCHAR(60) in arabic script, and user should ask for an ORDER BY french or by arabic.
In many SP, I'll use select ... where arabic_name ...
That's why specific collation is mandatory.
But I don't think that these structures will be modified, once defined. I've tried to deploy DDT packages on such a database (one column in one table with an arabic_collation), and I can modify all but that column. That's OK for me today (most important is to understand how to do with browsedialog in french AND arabic , I'll ask you later)