I don't think this possible, but I'm gonna go ahead and ask anyways:

Is it possible to have a decimal dimension, but with the tolerances as fractions?

I know it's an odd question and not recommended in most engineering circles. But it comes up occasionally here at my company. We like to use fractions for dimensions with looser tolerances. The problem comes up when we have things in integral inches. In that case, it gets changed to a 2-place decimal: i.e. '1' becomes '1.00'. The thing is that occasionally we add tolerance that stays a fraction, 1.00+1/16 /-0. I know it can be shown as 1+1/16 /-0 or 1.00+.06 /-.00. but that's not what I would like.

Any thoughts?

Can you please say it in more details? Or maybe give some samples.