Following script should make it clear in regards to Question 1.

    declare @i int = 0;
    declare @temp table( id int, latin varchar(1) collate SQL_Latin1_General_CP1_CI_AS, jap varchar(1) collate Japanese_90_CI_AS )
    while (@i < 256)
    begin
    insert into @temp 
    select @i, char(@i)  collate SQL_Latin1_General_CP1_CI_AS ,char(@i)  collate Japanese_90_CI_AS  
    
    set @i = @i+1;
    end
    
    select * from @temp
    order by latin
    
    select * from @temp
    order by jap

2. Collation is the rules on how to sort [have a read on this][1]. If your columns are `nvarchars` instead of `varchars` answer becomes different.

3. Any answer would be opinion based, most of companies above don't use single type of database (graph, bigdata, etc.), I never had a need to use anything but `SQL_Latin1_General_CP1_CI_AS`. (I never had to work out of Europe ether)

4. If you are using `nvarchars` = you are using `unicode` and unicode is like Chuck Norris - covers everything (twice).


  [1]: https://msdn.microsoft.com/en-GB/library/ms143726.aspx#Code_Page_Defn