Re the combination of output and input value calc's. If I understand the latest discussions it is possible to create a schema will allows: 1. All input fields to be calculated. 2. All output fields to be calculated and have different values than the input fields. #1 means that any file, including an empty one, can be used for input. #2 means that although the user will interact with the input fields none of the input fields will Ever appear in the augmented infoset or output file. It also means that no infoset value that is set by a user will ever appear in the output file. Sounds like this is pretty well hashed out. By the way, I'm not using the above contrived, unrealistic example as criticism. Rather to Illustrate that it looks like you've provided closure to the issue and haven't left any 'gotchas'. Of course, now you've got me thinking about adding a random number function or two so I can write a schema for test data generation! (ok, maybe that last one was a bit sarcastic)