Re: The Optimal Number of Gyroscopes
The optimal number is, of course, three because if one disagrees with the two others its results can be ignored in favor of the most common result.
Independantly of whatever argument there may be about using two, I think the point is that one of the three has gone bad, two are left and you need to have a spare. So they're shutting down the faulty one, keeping one as a spare and only using one gyro. When that one fails, the spare will pick up the job and Hubble will continue to be useful.
That's how I see it anyway.