The very well documented, very significant health impacts of obesity, need to be compared with any possible significant health impacts of dieting.

The problem here is that most fat people don't eat healthily, and don't exercise enough. Even if there are people for whom such a 'constant weight' threshold does not exist (or would otherwise lead to other health issues), these people do not constitute the bulk of the problem. By trying to explain every single case, one loses the most important part of the message.

To this person, my proposal would be to measure your weight, and reduce your caloric intake until your weight is constant. Such a feedback loop will ensure that you don't gain weight. This means that eating in the traditional diet range is the solution. The article does not bring any data disputing this, and instead hints at the fact that some scientist investigate whether other factors might also play a role. BTW my suggestion isn't anything new, just do a google search for the 'Hacker's Diet'.

Dieting cannot be thought of as a short term thing. Keeping your weight under control means not eating too much (calorie-wise) for the rest of your life.

For the record I'm not in favor of starvation. I'm just using an extreme case to remind everyone here that basic conservation laws also apply to human beings. No calories can ever be 'created', and even in 'starvation mode' there's no way that in the long term you will gain weight if you spend more calories than you consume. Energy conservation provides a strong limit for the importance all other effects that play a role in weight gain/loss.

There are indeed many factors playing a role in weight gain/loss. I have yet to see evidence that shows that the difference between the energy input and energy use is not the main one.

Contrary to what you're saying, the problem has everything to do with thermodynamics. You just don't seem to get the "simple thermodynamics" argument.

You're right that the energy used by the body might vary, and right that it's difficult to control or assess the difference between the input energy and the used energy. Still, you're missing an important limit in your calculation.

At 0 calories input, your body may well decide to do the following:
- Use only 750 calories
- store -750 calories
- poop 0 calories.

There is no way around it. You can *force* your body to use more than you eat, and you *will* lose weight. This can be done by eating less, or moving more. This isn't easy, would not necessarily be considered 'healthy', but that's not the point. The point isn't about whether you'll feel depressed or tired, it's about conservation of energy.