Here's an interesting piece on Body Mass Index - slightly polemical (and note the declaration of interest). Nonetheless it makes a number of interesting points e.g. if the definitions of 'overweight' in childhood are somehow linked to being in the top 5% of the age-for-weight distribution, how come more than 5% of children are said to be obese? There is an answer to that, but you'll need to read the article to find out.
BMI is so central to public health that we need to be aware of all the ins and outs of the arguments.