Infrastructure

If we were to use a single token to represent all of a person's identities in the log, then we would lose information regarding which of those identities they used when accessing a service. The DataONE identity_manager keeps track of which user identities are equivalent, so its a straightforward task to post-process a log to collapse all equivalent identities onto a single identity if that is needed for a report or summary. Thus, we've concluded that its probably harmful to do this during the logging operation itself. Thus, these are needs for tools for post-processing logs in various ways (e.g., for identitiy equivalence here, or for identity anonymization as described in ticket #1872), so I am converting this ticket to a feature request for a new tool, rather than as a change in how logging is conducted per se.