In the ancient and sprawling tangle of databases at my workplace, columns that are used for case-insensitive search sometimes have their collation set to SQL_Polish_CP1250_CI_AS, and sometimes to Polish_CI_AS. Sometimes those match the database-wide collation of SQL_Polish_CP1250_CS_AS or Polish_CS_AS, sometimes not.

I know that both of these collations handle Polish fine (since different Poles recommend each as the case insensitive collation), that Microsoft's own documentation lists Polish_CI_AS as the default collation for Polish, and I should probably pick SQL_Polish_CP1250_CI_AS if the database-wide collation is SQL_Polish_CP1250_CS_AS...

...but I have no idea why should I pick one over the other in a fresh environment.

So: what's the difference between SQL_Polish_CP1250_CI_AS and Polish_CI_AS? Is there any practical difference for choosing one over the other? Are there any legacy-system-related considerations at play here?