This presentation describes legacy code page architectures and Unicode as they are actually implemented in client-server enterprise computing.

Client-server computing is essentially an attempt to overcome limitations of unreliability and low bandwidth in data communications networks.

However, in the real world client-server computing quickly becomes distributed computing. This raises the issue of what to do with transfer of text data between different legacy code pages on different operating system platforms.

The presentation briefly reviews the emergence of our current concept of code page architecture from earlier technologies for numerical encoding of text data. We review the major families of code pages for legacy systems.

There is one generic solution to the legacy code page problem, used in all implementations for enterprise computing – provision of a transcoding service. We discuss the characteristics and built-in problems with transcoding services.

Transcoding services do not solve a second problem - combining unrelated scripts (writing systems). Unicode solves this problem, but the solution is rarely needed in real commercial applications. However, Unicode is of immense value in interfacing enterprise client-server with emerging technologies.

Finally we show two ways Unicode can be applied to client-server. Unicode can substitute for and eliminate transcoding. However, in many implementations, Unicode must be used as a supplement (“just another code page”) because of the importance of real data stored in legacy code page systems.