> Um no. It's a significant problem. Once you get beyond the top 1% > or so, It's not data mining, it's text mining (twisting David Evans's > concept a bit) Could you explain that a bit further? Beyond the top 1% keywords? Peter