Microsoft SQL Server (MS SQLS) has supported Unicode for the last two verions, each time with different ideas of how text is to be stored and collated. The purpose of this paper is top describe what MS SQLS does in order to support the storing, normalizing, and sorting of Unicode data, both at the engine level and at the level where you have some control over the ordering. Many of the specific issues in MS SQLS 2000's new COLLATE keyword will be addressed, including both the features and the limitations of this powerful functionality. In addition, the mixed meaning of SQLS collations and code page conversions will be discussed.

The conclusion of the paper is that MS SQL Server can be a very compelling multilingual platform for its Unicode support, but with some important caveats that must be understood in order to properly make use of its capabilities.