MetaData needs to maintain a hierarchical ancestry of Lineage. All ESRI operations (GeoProcessing, Export, Copy, etc) should produce Lineage which inherits from all input source(s). This would include a few key tags (Abstract, Coordinate System, Author, Creation Date, etc) as well as inheriting any hierarchical lineage already present in the input from previous processing steps. This also requires a more robust MetaData Viewer that can navigate the hierarchical tree of lineage/ancestry. And the standard for storing this hierarchy should be documented and published for consistent use by 3rd party extension and data suppliers.
Value: we hope it is apparent how useful this information would be to fully understand how a dataset was derived from various other inputs and processes. And for the individuals performing the processing, how useful it would be to remember what was done when it needs to be revised or repeated with new data.