The current way that incorrectly named configuration parameters are handled (they are silently ignored) seems to be a poor user experience for novice node operators. Because config values are only fetched when needed at runtime, if a config parameter is passed in with a typo, it is assumed not to exist and the default value is used.
Wouldn't it be an improvement to add logic during node initialization that iterates through all user-supplied configuration parameters and checks that they are supported parameter names? It seems like this would require the creation of a single data structure that can act as a directory for all available parameters.