Don, I agree there are practical limits, but they are not completely impractical either, you probably understand what I am getting at.
Zach and Don : Please understand I am not trying to make a large issue of a small change, I simply do what I can to bring up edge cases (usually I fix them, but this isn't a piece of software), because edge cases can not be ignored. They need to be addressed, sometimes this can be quick, easy and painless, other times, it's hard and painful. It varies.
To all, the reason I bring up these edge cases, is because we're pushed to try and adopt uniformity and singularity in a fragmented system (See Android as a resemblance to how divisional training is conducted, some ARTCC training programs are jelly bean, others are gingerbread) and there are 20 other variations. The ultimate goal should be to get all training systems to Jelly bean, because once that happens, these discussions won't need to be had. Because the uniformity will already be there.
How (and whether) that is implemented is not my call, but know I will do my best to help that cause.
P.S. Don, you are correct in that we should never force control, but we should provide it if requested. And your representation of the soft bounds was perfect in my opinion. Just a note I thought should be made a point.
Cheers!
Rahul