I would like to create a repair utility that my end users can use with little or no direction from me. Most of my end users are just users, and not IT people -- they are insurance agents or office workers.

Helen's book describes eight steps to data recovery. Some of those steps have options that can be exercised in a worst-case situation. For example, checksum errors can be ignored while validating. So, when creating a GUI front end for recovery, is it safe to always assume the worst-case?