The Cold Dark Matter (CDM) model has an elegant simplicitly which makes it very predictive, but when its parameters are fixed at their `canonical' values its predictions are in conflict with observational data. There is, however, much leeway in the initial conditions within the CDM framework. We advocate a re-examination of the CDM model, taking into account modest variation of parameters from their canonical values. We find that CDM models with $n=0.8$--0.9 and $h=0.45$--0.50 can fit the available data. Our ``best fit'' CDM model has $n=0.9$, $h=0.45$ and $C_2^{T}/C_2^{S}=0.7$. We discuss the current state of observations which could definitely rule out this model.

6 pages, self-unpacking uuencoded compressed postscript, with 4 included figures, to appear in MNRAS. References added and Fig.3 corrected. Also available at http://physics7.berkeley.edu/cmbserve/gen.html