This topic applies only to the IBM Business Automation Workflow Advanced configuration.

Working with generalized flow activities

Edit online
Draft comment:
This topic only applies to BAW, and is located in the BAW repository. Last updated on 2025-03-13 12:15
Draft comment:
This topic was viewed 8 times since its publication
The generalized flow activity is very similar to a parallel activity in that you can nest other process activities within it, and then control the execution order of those activities through links. The generalized flow activity differs however in its ability to use conditional links to loop back to previous activities in the sequence.

About this task

The generalized flow and parallel activities can both be used to model a business process in a graph-oriented manner. The main differences are:
Table 1. The main differences between generalized flow and parallel activities.
Type of activity Supports parallelism? Cycles? Fault links
Parallel activity Yes No No
Generalized flow activity Yes Yes Yes

You can create a generalized flow activity as follows:

Procedure

  1. Drop a generalized flow activity onto the canvas.
  2. Populate the generalized flow activity with the necessary activities.
    It is important that a generalized flow has:
    • exactly one start activity (an activity with outgoing links but no incoming links)
    • one or more end activities (an end activity is an activity with incoming links and no outgoing links).
  3. Link the activities together as follows:
    1. For each element (action or structure) that forms part of the generalized flow, click the element to select it, right-click to launch a menu of choices and select Add a link. Move the mouse so that the free end of the new link is positioned on the target element and click.
      A link is created from the first element to the second.
      Note: On the same menu as the Add a link option you will also see Add a fault link which is tool for creating simple fault handling behavior in your generalized flow.
    2. Create the necessary links between each of the nested activities. When you are done, there should be a clear control path through the activities.
      Note: Crossing links (links where the source activity and the target activity reside in different structured activities) are not supported within a generalized flow activity.
  4. Optional: Create fault links where appropriate.
    When faults occur in your business process, fault handlers are typically engaged to deal with the fault. The generalized flow activity offers a simplified fault handling procedure. From any scope or basic activity (excluding the Throw and Rethrow activities) you can add one or more fault links. A fault link in an activity is followed if the specified fault occurs while the activity is running. You can define catch fault links for various conditions, or you can create a catch all fault link, which will be followed should any fault occur that is not covered by a catch fault link. If multiple fault links are modeled for the same activity, best match decides which fault link will be followed. The fault catching rules are the same as for fault handlers.
    Note: No more than one fault link can catch an occurred fault.

    You can terminate the fault link at any activity within the generalized flow - for example you may choose to direct the fault link to a terminate activity to have the business process stop if a fault occurs, or you may want the business process to skip the rest of the steps in the generalized flow and exit to the next activity in which case you would terminate the fault link at the last activity in the generalized flow. You can also use a fault link to create a cycle and loop back to a previous activity.

    To create a fault link, right-click the activity and select Add a fault link. Drag the cursor out over the canvas. A black line is now attached to the cursor, and when you hover over a valid incoming link icon or valid gateway, the crossed out circle disappears, and the link snaps into place. Click the activity or gateway that you want to link to. The fault link becomes a striped red line and connects the two activities.

  5. Specify the diverging gateway type.
    A diverging gateway is made available for an activity when two or more links (excluding fault links) start at this activity. You can specify one of three types of diverging gateway.
    • Screen capture of a Split diverging gateway.
      Split - exclusive pathway. When the activity runs successfully, the first link that evaluates to true is followed, at least one of the links must evaluate to true. None of the other links are followed. Since the default condition for a link is true, it may then be necessary to adjust the execution order to make sure that the appropriate link is fired. Links starting at this gateway type can loop back to a previous activity in the flow, so these links can create a cycle.
    • Screen capture of a Fork diverging gateway. Fork - parallel pathways. When the activity runs successfully, all links are followed. Link conditions cannot be specified. If this activity is skipped, all links are followed with a false flag, which influences the behavior at the converging gateway.
    • Screen capture of an Inclusive OR diverging gateway. Inclusive OR - parallel pathways with conditions. When the activity runs successfully, all links that evaluate to true are followed with a true flag, at least one of the links must evaluate to true. All links that evaluate to false are followed with a false flag. When the activity is skipped all links are followed with a false flag.
  6. Specify the converging gateway type.
    A converging gateway is made available for an activity when two or more links (including fault links) end at this activity. You can specify one of three types of converging gateway.
    • Screen capture of a Merge converging gateway.
      Merge - exclusive pathway. When the first link is followed with a true flag, the gateway is evaluated and the activity commences. No synchronization happens with the remaining incoming links.
    • Screen capture of a Join converging gateway. Join - parallel pathways. All incoming links are synchronized, the gateway is evaluated when all links are followed. When all links are followed with a true flag the activity commences. When all links are followed with a false flag the activity is skipped. When some links are followed with a true flag but some with a false flag the runtime throws an exception.
      Note: In cases where this exception can happen, a warning is displayed while modeling the process or when deploying the process.
    • Screen capture of an Inclusive OR converging gateway. Inclusive OR - exclusive or parallel pathways. This gateway can be used to merge exclusive paths, or to join parallel paths, or both. When merging exclusive paths, the gateway is evaluated and the activity commences when the first link is followed with a true flag. When joining parallel paths, all incoming links are synchronized, the gateway is evaluated when all links are followed. When one or more links are followed with a true flag the activity commences. When all links are followed with a false flag the activity is skipped. When exclusive and parallel paths converge at this gateway, it is evaluated in the reverse order of the opposite converging gateways. See troubleshooting for more information.
  7. Create any necessary conditions as follows:
    1. Click a link to highlight it.
      Links commencing at a fork gateway cannot have a condition applied.
    2. In the Details page of the properties area, click Create a new condition.
    3. Configure the condition.
      You can use a Simple expression language (True, False, or Otherwise) or program the condition using either Java or XPath. If you choose Java™, you can use the snippet editor to visually compose the code, or enter it directly into the Java editor yourself.
    4. If necessary, adjust the link evaluation order by selecting the Split gateway with multiple outgoing links, and clicking the Link Evaluation Order tab in the Properties area.
  8. Create conditions for you fault links as follows:
    1. Click a fault link to highlight it.
    2. In the Details page of the properties area, choose Catch or Catch All.

      For a catch fault link you can specify the fault name to be caught. This can either be a built-in or a user-defined fault. Additionally you can select a fault variable. The fault variable must be declared on one of the parent scopes or the process scope. Default variable scoping rules apply. For a catch fault link either the fault name, the fault variable or both must be specified. A catch all fault link will be followed if a fault occurs and no catch fault link is applicable. If you specify only catch fault links and do not cover every eventuality with your conditions you may end up in a situation where a fault occurs but the process cannot follow any of the fault links. In this case the fault will be thrown to the next scope or to the process scope where normal fault handling continues.

      If more than one catch fault link is specified for an activity, the following rules determine which fault link is navigated (these rules are the same as for a fault handler):
      • If the fault has no associated fault data, a catch fault link with matching fault name will be selected (providing the catch fault link has no variable assigned to it). Otherwise, a catch all fault link is chosen when modeled.
      • If the fault does have fault data associated with it, then a catch fault link with matching fault name and variable data type will be selected. If there is no fault name specified, then a catch fault link with a matching data type will be selected. Otherwise, a catch all fault link is chosen when modeled.

Example

Here is an example of a generalized flow activity that processes an order for a color copy that comes in over the internet.
A generalized flow activity that processes the order for a color photocopy.

In this example, the order of a color copy comes in through the Receive copy order activity. A human task is then assigned to an employee in the Set up print activity who sets up the source file for printing. Part of this set up process includes adjusting the colors so that they are a close match to the original. When this set up process is complete, the flow continues to the Print activity which produces the copy. The output is then directed to another human task, the Color inspection activity. This employee has the responsibility to decide if the colors are a close match to the original. If they are not, then the evaluation is set to false, and the flow is returned to the Set up print activity. This loop will continue until such time that the employee at the Color inspection activity decides that the color is acceptable, and the process is allowed to process to the Ship copy activity.