Michael Reichmann, in his "Understanding DSLR Workflow" article on this site, states:
"Digital has a remarkable ability to extract detail from the shadows. When shooting digital, always ere (if you have to) on the side of underexposure."
whereas Bruce Fraser - another highly respected digital processing expert states (in Real World Camera Raw):
"correct exposure in the digital realm means keeping the highlights as close to blowing out, without actually doing so, as possible.....it's better to err on the side of slight overexposure".
So who is right?
I have to say the explanation given by Bruce Fraser seems quite convincing - its too long to go into here but basically there are more levels of data at the brighter end of the scale, so if you subsequently have to "stretch" the data in levels or curves then you'll get more data to play with.