To correctly process bidirectional text, applications must assume that text is stored in computer memory in a "different way" from when presented on an output device. In other words, a text-conversion process of some sort takes place when bidirectional text is retrieved from storage (files, memory) to be displayed on the screen (or printer), or when it is read from the keyboard and stored.
There are many ways to represent (store in memory, or files) bidirectional text that contains Arabic or Hebrew characters, English characters and numbers (digits). Different applications use different methods to store their text data and use different conversions before displaying their text.
Applications may choose to use different formats to represent bidirectional text, depending on the nature of the processing that they perform on this text.
In order to support bidirectional text processing, the PM support for Hebrew and Arabic defines a set of parameters that describe the different "features" that any Arabic or Hebrew bidirectional text may have. These parameters are called "Bidirectional Attributes".
A bidirectional application can describe the nature of the bidirectional text data it processes, by identifying a set of bidirectional attributes. These bidirectional attributes, which are implemented in PM in a hierarchical model, uniquely define the bidirectional text and the conversions it must go through, to the PM system and to the PM bidirectional services.
Below is the list of the bidirectional attributes as well as a short explanation of each one of them.