> I want to build a high accuracy low range time of flight laser range
> finder. ... I'm not sure which is harder to get a
> pulse back from - a mirror or a perfectly non-reflecting object).
Is there some pressing reason to use a single pulse as your light input? If you operate a CW laser with superimposed modulation, then FM that modulation, the reflected light will have the frequency of your laser from 2*r/c seconds ago, and you can beat that (mix, using normal RF mixer) against the source to generate a low-frequency difference. DSP tools can easily measure the difference and there's your distance!
The laser and reflection modulation is narrow-band, and the mixer isn't sensitive to out-of-band signals, so signal/noise performance should be good.