What's really happening is perverse incentive structures. Ideally all energy use of every app would be accounted for and, somehow, relayed back to the original "responsible agent" (e.g. a signed app being responsible for all its subprocesses). This then needs to be tied to some valuation.. in an app store, you could say e.g. lower commission based on energy efficiency, or just a higher ranking score.
Currently, resource usage suffers from the tragedy of the commons, and it will remain that way until we fix the incentives.
OSX is starting on that road, as you can see with the Energy tab in activity monitor. I wonder if they will take further steps with it, but it's a good start. (You'd then want e.g. browsers to allow push the accounting down to invidual sites; now we're suddenly attacking death-by-JS as well.)
One day! But until that day, you are right: nobody tests those things. Why would they? Honestly, if you're creating drivers for something, worrying about energy use is... is so inconsequential it's borderline altruistic. Who will ever, really notice? It's basically charity work. That needs to change.
There are also orders of magnitude difference between a sleeping device and an idle device, and an idle device to a fully loaded device. We are talking three weeks to 20 hours to 35 minutes. And that gap will only continue to get larger
The battery life on the X1 Carbon is about one and a half to two hours, which i think is acceptable given that the battery is reporting less than half of its original capacity.
I tend to go with older laptops, but they are definitely fast enough and I like how robust the experience is.
That's evidence that there exists something that Linux users recommend to people who complain about having battery problems on Linux, not evidence of widespread problems with Linux that require expertise to get the operating system to "work" at all.
Of course Windows is shit in this respect too these days. I got two extra hours of battery life from my X1 Carbon by wiping the disk and installing Windows LTSB. But even then I still have to police things, watching for driver updates or software installs that randomly leave some service running that keeps the CPU from sleeping. It’s like nobody tests these things.