If you set the coordinate system when creating a new feature class or dataset, we use the coordinate system's (Map Projection) Valid extent. this is larger than the coordinate system's valid area of use. for example, in your case, the extent of North Carolina Lina is the valid area of use. the 'valid extent' is what's supported mathematically by the lambert conformal conic projection which covers a much larger area.
For geographic coordinate systems, the valid extent is greater than the world.
We usually set the minimum extent/domain and resolution values. That often means that the maximum values are much larger than required. This cocould happen prior to 9.2 too, but wasn't as obvious.
There are several reasons we do this. one is that we have the information available about the valid extent available now. two, with "High Precision" aka 64bit/53bit, we can support the valid extent and also maintain reasonable resolution values. if you use the actual area of use of your data in a high precision GDB, the resolution value will be ridiculously small, which can cause performance problems.
You may have seen the 2147483647 value before in connection with the pre-9.2 Geodatabase? Values, after an internal conversion using the domain and precision (1/resolution), had to fit between 0 and 2147483647. at 9.2, we now use 2 ^ 53-2 or 9007199254740990 for the max permitted value. we have a lot more space to work.
There's going to be a White Paper on this, in a month or so.
So, the things to remember at 9.2 when setting up a feature class/dataset are:
1. Set the Coordinate System
2. The tolerance values shocould reflect the accuracy of your data.
2. The resolution values shocould be at least 10 times smaller than the corresponding tolerance values. The resolution values dictate the minimum distance between two points in the feature class.