Analyze Object Links Recommended Threshold and Business Critical Objects Properties
In LiveCompare 2.6 and earlier, the Analyze Object Links action produced an optimal set of executables, which when tested would exercise each of the transport objects provided in the action’s Source table. In this release of LiveCompare, the Analyze Object Links action includes a Recommended Threshold parameter, which specifies a minimum percentage of impacted executables and screens that should be recommended for testing. This parameter may be used to augment the list of most-at-risk executables and screens with used objects specified in the action’s Usage Counts table. Added objects are taken from the set of impacted executables and screens. An object will not be added if it is used but not impacted.
The Analyze Object Links action also includes an optional Business Critical Objects parameter. If this parameter is used it should contain a set of objects, each of which will be recommended for testing if it is impacted by any of the Source objects provided to the Analyze Object Links action.
The recommended objects for testing are stored in the Analyze Object Links action’s Result dataset. This dataset is populated as follows:
- Any impacted business critical objects are written to the Result dataset. If they are used, their usage count is set to the COUNT value in the Analyze Object Links action’s Usage Count table. Otherwise, their usage count is set to 1.
- An algorithm is run to determine a minimal set of covering executables and screens, which when tested will exercise each of the source objects. These executables are written to the Result dataset.
- If the Recommended Threshold has not been reached, the algorithm is run again, having removed the executables and screens that have been recommended so far. This provides the next best executables and screens to recommend, and a new set of covering executables and screens.
- Step 3 is repeated until the Recommended Threshold has been reached.
If the Recommended Threshold value is set to 0, the Result dataset will contain an optimal set of most-at-risk executables which are recommended for testing, as well as any impacted business critical objects.