We can debate the morality of energy efficiency as long as we want, but whether we like it or not eventually it comes down to cost-effectiveness. Does a windmill or solar panel produce enough electricity to make it useful in real life? In some situations, yes. In others, no.

For software, does bad code in one application cause an increase in 5 watts of power? 10 watts? 100 watts? What does it cost in hourly wages to have a developer check for such efficiency? How does that cost compare to the cost of a few extra watts? And more, what is the cost of implementing the controls that would be required to check whether or not efficient code is affecting the bottom line in the server center? Do we hire a consultant? Do we increase the QA staff? Is extra equipment required? Do we purchase a special software package, that itself consumes power, to monitor power consumption in other applications?

Maybe the power consumption of individual work stations should be considered? One wasted watt spread over 10 million machines adds up. On the other hand, it often adds up for someone else; not the writers of the errant app.

This is a very interesting point and may be something we all watch someday. However, something tells me it will be a long time before there is widespread concern about it.