From a system point of view, one can argue that sort order is a display
function. Systems that are designed to display data in sequence (as index
listings, classification-order listings etc.) need to have consistent data to
work with and rules to apply. Mechanisms to achieve that consistency in data
are surely a proper part of the rules for the data framework. Predictable
sequence is one of the tools for assessing and using the data (I consistently
use name index lists to discover and resolve apparent discrepancies,
distinctions and duplications in forms of name).
If not, we're going to be taking a giant step backwards!
[log in to unmask]