From time to time, we come across prospects and customers that want to be able to upgrade a ThinLinc installations without incurring downtime. It would be nice if our upgrades could be applied to running systems without fearing for incompatible changes, without having to allocate service windows, and without having to perform an "atomic" upgrade of the entire cluster at once. Use cases for this include when ThinLinc is used inside operations centers, hospitals, or any other case where staff are working 24/7. Related: bug 7112.