Thanks, Jason, but - unless I'm missing something - neither of these does what the instigators of XBRL had in mind. Namely, to give anyone the ability to freely and easily consume the financial statements in the EDGAR database. In other words, to construct a model in a spreadsheet with pointers to a specific taxonomic element (or combination of elements) within a filing for a defined time dimension (a period or date) along with the ability to derive values (for example, the ratios, margins or the sort of triangulations between financial statements you describe). Today, if I want to pull numbers from EDGAR, I'm pretty much doing what I did 20 years ago: copying and pasting from an HTML document into a spreadsheet where the main value is ensuring the numbers in my model are exactly what was published. This is still time consuming an tedious.