I'm looking at Torque, looking at its db schema input files, and thinking: couldn't I create a system where I used Torque's classes to deal with mapping database fields to a schema (by schema I mean also mapping the db fields to their semantics: what their possible values are, essentially - the stuff that could be put in DB rules, but usually ends up in application code), and then never generate business objects but instead use Torque to help me keep track of the schema and the fields? I want to build a dynamic data retrieval application (posted a similar question on OO/Patterns/Refactoring, but just had the thought re: Torque). . . the set of fields in the database is going to change, and I'd rather set up a generic mechanism to deal with them, rather than explicitly changing my app's UI. Guess I'm looking for anyone who's used/thought of using Torque (or anything else, for that matter) in that kind of way.