The problem with etiologially associating smoking tobacco with death is that it can't be simply demonstrated that a dose-effect relationship can be measured and all such associations must be studied at great length and subjected to rigorous control methodologies to arrive at a statistically significant cause-effect relationship.
Long-term studies clearly show that career-smoking is harmful to one's health but so are many other normal and daily acivities such as driving in bumper-to-bumper traffic or crossing a busy street.
What hasn't been shown is a clear and present danger to being exposed to ETS in ordinary living arrangements.
As I said, I am opposed to general bans on smoking by government because it amounts to another abridgement of free association among groups. Current bans prohibit a gathering of all-smoking patrons in any place where one or more of the participants is an employee; only an idiot would think that this would lead to an improvement of these people's health - they will continue to smoke even if elsewhere, singly or in casual groups, and the general welfare neither gains nor loses.
It is the overarching reach of the big hand of government to which I object.