On Fri, Dec 30, 2005 at 08:58:03AM -0800, Jim Busser wrote: > Well, does that mean when someone has vacated their exam room or > desk, and that is a shared work space, the expectation is that when > the previous person leaves and the new worker (doctor or secretary) > sits down at that machine, not only must the GNUmed client be logged > out and back in, but the user must *also* change themselves to be a > different user account on the machine or network?
Actually yes. Some countries even put up regulations for that (US: HIPAA). They mandate auto-logout with a timeout. However, in many cases that won't really be practical in a smaller setting such as a GP practice where it may make more sense to have one dedicated *system* account (gmuser, perhaps) which the machine is logged into. Users are then only expected to logon to the client with their own *db* credentials. This does not really, however, provide the amount of security that would be needed. This can be overcome by auto-logging out the system account after a timeout and do re-login via fingerprint. It may be worthwhile some day to add "change credentials" to the GNUmed client proper which would then allow timed auto-locking the client (with auto-save, then, perhaps) and unlocking with different credentials. Or changing credentials on the fly upon invoking a menu item. Using your own account (system or db level or both) might be made attractive by somehow measuring activity (level, not content) and using that as a goodwill currency in, say, acknowledgements or even payouts. So, the entire thing is more social engineering than anything else. We need to provide the appropriate tools for that. That's also why PostgreSQL access configuration seems very complex. Because sites have very diverging needs. Karsten -- GPG key ID E4071346 @ wwwkeys.pgp.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346 _______________________________________________ Gnumed-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnumed-devel
