On the one hand, double stars are a nuisance for data reductions. On the other hand, they are scientifically interesting and important. After all, binaries probably constitute the majority of the overall stellar population. Thus they are crucial e.g. for our understanding of the formation of stars and planets in general. The Gaia mission sees, discovers, measures and parameterizes double stars - both optical pairs and physical binaries - in a surprising multitude of ways. Each of these ways poses an operational challenge as well as a scientific chance. Once fully exploited they will give a strongly revised picture of stellar binarity statistics. And they will remove all the disturbances caused by duplicity in the astrometric and photometric data of Gaia DR2. Gaia DR1 achieved an effective angular resolution (i.e.pair separations) of 2 arcsec, DR2 of 0.4 arcsec. But the actual optical resolution of the Gaia instrument is about 0.15 arcsec, and there are ways to detect and measure pairs down to the milli-arcsec level. All this can be done for hundreds of millions of stars, but it means a few more years of hard work by the Gaia data reduction consortium.