For instance, there are German collations for dBASE, for Paradox, and an InterBase specific collation “DE_de”. InterBase defines several collations that are “the same, but different”.
InterBase duplicated known bugs in dBASE/Paradox collations to preserve compatibility. An important design decision was that InterBase collations must match the corresponding dBASE/Paradox collation PRECISELY, to make them compatible. The collations were re-implemented in InterBase and validated against the original product. Many of the International collations supported in InterBase were originally designed for Borland’s Paradox and/or dBASE product lines. These examples should also work for any InterBase 4.x or 5.x version – as the architecture has not changed since originally implemented for InterBase 4.0.
This document describes how to create drivers in the InterBase Internationalization architecture. Unlike the 3.3 architecture, InterBase 4.0 used the then-standard SQL-92 Internationalization architecture. InterBase 4.0 built on the foundation established by v3.3 and v3.2J and combined both single-byte and multi-byte character set support into a single product. There was no ability to extend the functionality of the international parts of the product. It only supported codepoint order collation. The 3.2J version supported Japanese data stored in EUC and SJIS formats. Separately, a version 3.2J (Japanese) was released in 1992. Version 3.3 also supported the ability to extend the International functionality by adding additional character set or collation modules. Only single-byte character sets and European-style collations were supported at that time. In June 1992, InterBase v3.3 was released, which included the first support for multiple character sets in a database and alternate collations for text. Only binary collation order was supported for CHARACTER(n) data. It was up to the application programmer to know what character set was used to store the data, and how to convert between different character sets. Prior to InterBase v3.3, InterBase defined CHARACTER(n) data as simple 8-bit bytes. To download the above as a (30.5kb) (.zip) file
This document uses international characters for illustration. An example using InterBase 6.0 is included.
This paper describes how to create a custom collation or character set driver for InterBase.