We have seen that, despite the intrinsic complexity of the dust opacity in external galaxies, in certain cases the problem of dust obscuration is treatable. A 'recipe', under the name of attenuation curve, is available for the reddening correction of the integrated light from star-forming regions. The strength of the starburst attenuation curve is that its derivation is purely empirical and does not rely on models. The curve is therefore applicable at least to the class of objects it has been derived from: the central star-forming regions of galaxies.
One of the open problems is understanding the limits of applicability of the curve. In the case of isolated HII regions, where the SF processes are less energetic than in the case of starbursts, the dust can survive the less harsh environment and be uniformly distributed with the stars. For this reason, the attenuation curve will not be generally applicable to HII regions. Similar arguments can be used in the case of 'quiescent' galaxies.
Other open problems are the meaning of Equation (3), namely, the discrepancy in global attenuation between stars and gas though they are still correlated, and the physical meaning, if any, of the effective color excess E(B - V)s of the stellar continuum. These and other issues will require further investigation.
The author acknowledges valuable discussions with Tim Heckman, Max Pettini, Gerhardt Meurer, Mauro Giavalisco, and Claus Leitherer.