It is often stated that Gaia will provide a 3-dimensional view of the Milky Way via the measurement of distances. However Gaia (of course) measures angles with finite precision and not direct distances. In particular it is estimated that 80% of the final Gaia catalogue will have parallax measurement significance at the 5-sigma level or less. At such levels of precision the non-linear reciprocal relationship between parallax and distance results in some obvious (how do we handle negative parallaxes?!), and not so obvious (how do we assign confidence intervals?), problems. In a series of papers Bailer-Jones has provided a thorough analysis of these issues and presents Bayesian inference methods as a way to extract maximum information from the Gaia catalogue. This talk will provide an overview of these aspects, will illustrate with examples from TGAS in Gaia DR1 and will also give some general advice on how to make the most of the astrometry.