We present a luminosity function for similar to 4000 stars similar to 3-5 half-mass radii from the center of the globular cluster omega Cen (NGC 5139), based on images obtained with the Hubble Space Telescope's Planetary Camera as part of the Medium Deep Survey Key Project. The luminosity function (corrected for incompleteness) has a plateau at 5<I-M<7, and flattens out at M(I)>8.5. This flattening contradicts results from ground based data at the same radii, which imply increasing numbers of stars at faint magnitudes. We illustrate the pitfalls involved in converting luminosity functions to mass functions. With our current lack of knowledge of the mass-luminosity relation for low-mass, metal-poor stars, the shape of the mass function below m similar to 0.25 M. remains uncertain, even with a luminosity function reliably determined to M(I)=10 (m similar to 0.20 M.). We compare solar abundance, solar neighborhood luminosity functions with the omega Cen luminosity function, and find that they are indistinguishable over the range 5<M(I)<10. If this is not just a coincidence, then it suggests that either the stellar initial mass function over the range 0.2<ml.M.<0.8 is relatively independent of both environment and chemical abundance, or the derivative of the mass-luminosity relation in both V and I changes with metallicity in just such a way as to cancel any metallicity dependence in the luminosity function. We present a color-magnitude diagram for similar to 800 stars, which shows a well-defined main-sequence extending, to I approximate to 26 and four white dwarfs with 24<I<25. About a dozen candidates for equal mass binaries suggest an upper limit on the equal mass binary fraction of similar to 5% in the range 0.4-0.7 M..