To use Amlib with Unicode (e.g. Russian Characters not found in Windows ANSI Character Set) you will need MS SQLServer 7 with SP2 or higher (may also work with Oracle but not tested yet). However, SQLBase does not support Unicode.

AMLIB is currently installed using SQL Server at the Kazakhstan State Agrarian University (formerly part of the Soviet Union) using the Russian (Cyrillic) character set within Unicode. We are also investigating entry of CJK script characters into the SQL Server database with Amlib, which may depend configuration and code page of the Windows PC entering the data.

SQL Server now supports Unicode data types. which makes it easier to store data in multiple languages within one database by eliminating the problem of converting characters and installing multiple code pages. Unicode stores character data using two bytes for each character rather than one byte. There are 65,536 different bit patterns in two bytes, so Unicode can use one standard set of bit patterns to encode each character in all languages, including languages such as Russian that have large numbers of characters.

However, Unicode data needs twice as much storage space but is offset by the elimination of the need to convert extended characters between code pages. In SQL Server, the new data types that support Unicode are ntext, nchar, and nvarchar. They are exactly the same as text, char, and varchar, except for the wider range of characters supported and the restrictions on storage space (for instance, 4,000 characters instead of 8,000 characters).

To implement Unicode with Amlib:

• Select a correct SORT order when installing SQLServer – e.g. use Dictionary Order, case insensitive BUT DO NOT USE accent insensitive
• When installing SQLServer select Collation of General Unicode
• When installing SQLServer you will need to select the correct code page which will suit the most common character set used within the bibliographic records
• You need create the database and ensure all VARCHAR() datatypes are now NVARCHAR(), and all CHAR() datatypes are NCHAR()

For existing Amlib customers to use Unicode they will need to export their existing databases, install SQLServer with the correct SQLServer settings and Unicode data sets, and reload all existing data.