Small-scale inhomogeneities, or `clumping, in the winds of hot, massive stars are conventionally included in spectral analyses by assuming optically thin clumps. To reconcile investigations of different diagnostics using this microclumping technique,
very low mass-loss rates must be invoked for O stars. Recently it has been suggested that by using the microclumping approximation one may actually drastically underestimate the mass-loss rates. Here we demonstrate this, present a new, improved description of clumpy winds, and show how corresponding models, in a combined UV and optical analysis, can alleviate discrepancies between previously derived rates and those predicted by the line-driven wind theory. Furthermore, we show that the structures obtained in time-dependent, radiation-hydrodynamic simulations of the intrinsic line-driven instability of such winds, which are the basis to our current understanding of clumping, in their present-day form seem unable to provide a fully self-consistent, simultaneous fit to both UV and optical lines. The reasons for this are discussed.
