Sarah Jelinek wrote: > -The user starts 1000 AI installs via some scripts. None of the systems > being installed have a 'head' for the user to monitor the installation. > They rely on 'pinging' or some other mechanism to determine if the > install succeeded. In the case being proposed above, the user wouldn't > know why the system was 'hung', waiting for their input. And, they would > not likely be aware that AI went interactive without their specific > intervention. So, introducing interactive behavior is really unexpected > from the users point of view. > > One thought I had that we might consider is a way for the user to tell > AI to go 'interactive' in the event of a failure. That is, the default > behavior is always automatic, but in the case where users know they can > manage the systems being installed, allow them to enable this > interactive experience in the event of failures. We meet the > requirements for AI by doing this, I believe. > > Thoughts?
If I was managing 1,000 installs, I would hope there would be a central management console that would show me the status/progress of each and allow me to remotely, interactively deal with issues as they arise since I might not be physically located near any of the systems in question. Cheers, -- Shawn Walker
