There are basically two types of observations that should constrain dynamo models for the sun. One is observations of the sun's magnetic field in all its manifestations, which dynamo models should eventually predict accurately; the other is observations of solar velocities and thermodynamic structure and variations, which constrain the inductive processes available for the dynamo. With respect to magnetic properties, the level of detail in current observations of fields and inferred field patterns far outstrips current dynamo capabilities to reproduce, but models have moved beyond comparison to just the 'classical' solar cycle characteristics such as Hale's polarity laws and the butterfly diagram. With respect to velocity constraints, much more is known now about differential rotation and meridional circulation in the convection zone and tachocline, than when early results from helioseismology overturned the prevailing alpha - omega dynamos of the 1970's that required angular velocity increasing inwards. Meridional circulation has become a particularly important component, since recent models show it can determine the solar cycle period correctly for observed flow speeds. By contrast, so-called giant cells, which have been particularly difficult to detect, may not be as relevant to the dynamo problem as they were once thought to be. For the future, improved inference from helioseismic analysis of structure, velocities and magnetic fields in the solar tachocline would be particularly valuable.