\[~linda.martin] please take note of SQALE as you review. From a practical standpoint I'm imagining one issue per file (on first use?) with cost incremented per use of the API. Usually we use linear-with-offset for things related to size, but fixing this will require a large initial investment in design and implementation of a different approach, and then incremental effort to address each use of the API.
Do you agree that this is appropriate? Or should we just assess one high cost per file?
