]> WARNING: This is not a valid XBRL taxonomy; it is only an experimental view of another possible way of specifying taxonomies that are tree structures. Emphasis is made on leveraging features of XML syntax for clarity and human readability. The XSLT used to generate this document is meant for a taxonomy that consists of a tree structure, which requires that 1) there is only one rollup per type, and 2) there is one root that doesn't have a rollup, and 3) no rollups ultimately point to themselves. Note that the default* attributes on the root element provide another level of "cleaning up". Where a value is not the default, it is explicitly provided inline (including label elements). Also note that all reference elements have been relegated to an "appendix", referred to individually by referenceID attributes where applicable. In this particular taxonomy, it's possible that element names are not unique, because the "removeDots" flag was set to 'yes' (default), when the stylesheet was applied to the original taxonomy document. Exactly how XBRL instance documents would then refer to this "taxonomy" is not yet clear - perhaps with something like XPointer and XPath expressions. Note that by using a tree structure, elements have an identity outside their name alone - their context within the tree. Thus, they need not have unique element names, if there's another way to reference them from instance documents. This would remove the need for the visually redundant parent.child naming convention, which itself does not guarantee uniqueness without potentially needing to arbitrarily change names in a taxonomy.