The Madden-Julian Oscillation (MJO), a slowly eastward propagating envelope of tropical convective activity, is unfamiliar to many climate scientists and virtually unknown to the public, despite its important role in weather and climate within and outside the tropics. It is of intense interest to tropical meteorologists, though, because it is not predicted by classical equatorial wave theory for a dry atmosphere. It is equally of interest to modelers because most operational general circulation models do not simulate anything like an MJO.  Many theories for the MJO exist, but recent research suggests that it may be fundamentally tied to moisture fluctuations and their interactions with moist convection and clouds.  As such, better understanding of the MJO offers the hope of increasing tropical weather predictability on time scales of 20-30 days and providing a stringent metric for the fidelity of cumulus parameterizations and convective feedbacks on climate change. We will discuss how satellite and surface remote sensing observations and GISS climate model simulations of observed MJO events are beginning to constrain the convection physics needed to simulate it, the role of clouds in driving it, and the work that remains to be done to capture the MJO in models while not degrading other aspects of tropical climate.