Chapter 8 Building an Executable BPM: Managing Data and Choreography
A variable describes a data container local to a process. It represents data that is important for the correct execution of the process for example, for routing decisions to be made or for constructing messages to be sent.
It allows the process to maintain data and process history based on the messages exchanged. The variable value is retrieved from messages received by the process and can be used for loops and conditional branching. It is mainly used to build the messages the process sends to its partners.
For more information about loops and conditional branching, see sections Loops (Executable BPM) and Parallel Branching (Executable BPM.
Variables have no symbol in the diagram, but have the following graphical representation in the Browser:
By default, a variable defined under a given level can be used by any activity defined at the same level or at sub-levels.
You cannot create external shortcuts for a variable, however you can replicate a variable.
The name and code of a variable should be unique within the parent scope (package, composite process, or model). A variable cannot share the same name and code as another variable that is defined at the same level or under a parent of the current level. However, two variables can share the same name when they belong to different composite processes contained in the same package.
Unlike the analysis model, data are not conveyed by flows in the executable model.
When you move a variable between models A copy of the variable stays in the source model, if the variable is used by processes in that model, as external shortcut of variables are not supported.
When you move a variable within the same model A shortcut of the variable stays in the source process or package, when the variable is used by some processes in that package or composite process.
| Copyright (C) 2007. Sybase Inc. All rights reserved. |
| |