Firstly, I  discuss the constraints on astrometry in general. I compare the strength and limitations of the different options: space based, ground based seeing limited, and adaptive optics observations.

Secondly, I present applications of astrometry from myself and my collaborators. For the most objects that also includes what can be learned from astrometry. Nearly always masses can be constrained with astrometry. Often the measurements are also useful for other properties, like distances and orbits.

I discuss following objects:

the potential cluster IRS13E in the Galactic Center, the 'gas cloud' G2 there, the nuclear cluster there, the globular cluster and stream of Palomar 5, the Magellanic clouds, and the Carina dwarf galaxy.