There is a problem with dialogs based on the class “Dialog” that are used e.g., in the context of RunBaseBatch classes such as “Automatic release of sales orders”. Because of the dynamic generation of control names for the contained elements on the dialog that is based on a total numbering through all elements that starts with the variable fields in the “Records to include” which are gathered from the saved query in the usage data, the control names on a dialog are unreliable.

This may cause the Task Guide to point to a completely different field than described in the step (see Yammer post with image). This behavior will unsettle or even frustrate users that are supposed to use the Task Guide for training or guiding. This could even make it harder for new users to accept the ERP.

This problem of the Task Guide respectively the Task Recorder also affects the RSAT for regression testing which is based on them. Automatic tests may fail just because of the last saved values/queries.

The idea to solve this problem is to change how control names on the dialogs are generated to gain more independence from dynamically generated fields. Since each field is an element of a DialogGroup, which also have unique identifiers, the control names of the fields could be composed of the parent group name and the number of elements within that group. So the first field in DialogGroup "Grp1" would be named "Grp1_Fld1_1" and the first field in DialogGroup "Grp2" would be named "Grp2_Fld1_1" instead of e.g. "Fld2_1".

This would make the control names of fields in one group independent from the number and names of fields in any other groups on the dialog.