I know that is the conventional approach but the reason I ask is that If my calculations are right, 500mm of ballscrew raised from 20 to 60 deg C gives a total expansion of around 0.25mm in the worst case - assuming there is no rise in temperature of the ballscrew mounting i.e. the mill table / Y axis structure is held at 20 deg.

(the coefficient of expansion of mild steel at around room temperatures being 11.1 µm/m -°C, rising to 17.3 for stainless)

I can easily allow for end float in the oldham coupling of twice that without compromising the integrity of the coupling.