Bob Wolven has an interesting piece in netConnect about cataloging. He mentions our approach to standards, among other things.
Perhaps worse, the kind of consensus we have demanded drives us toward complexity. Our libraries acquire a vast and wildly diverse set of resources, yet we insist on treating all of them by the same rules. We prize consistency over practicality. If some works, in some contexts, benefit from a precise transcription of statements of responsibility, or from detailed recording of pagination and illustrations, we apply those same principles to all. We apply the same level of subject analysis to the 20-page pamphlet and the 1000-page treatise. We do this not out of obduracy or short-sightedness, but because it's the only way we have found to build trust among what is, after all, a very large and diverse group. [In Search of a New Model - 1/15/2008 - netConnect]
We do sometimes treat standards activity as if the desired outcome were socially acceptable consensus. This has meant that we may allow optionality or discretion in how data is represented, or, for example, we may suggest that data go into notes. This may have been more acceptable when actual data exchange was not very frequent, or data was created for human display. However, as more of our services are supported by communicating applications, and as the volume and variety of data transfers increase, this approach is less useful. Think of how we want to process data for faceted display, or for clustering into works, or think about using data to manage flows into mass digitization or offsite storage where we want to track volumes through workflows. We want to make sure that the full intellectual effort that goes into description is available for re-use by applications.