I've rarely heard a weaker argument. We have DST, because where I live, it would be light by 3:30 in mid summer. With DST, that's an hour later, and there's a bit more light in the evening. That's a net positive. If every region had to pick optimal time zone, you'd have at least twice the current number of TZs.

"The first real experiments with daylight saving time began during World War I. On April 30, 1916, Germany and Austria implemented a one-hour clock shift as a way of conserving electricity needed for the war effort. The United Kingdom and several other European nations adopted daylight saving shortly thereafter, and the United States followed suit in 1918."

Personally, I hate the practice and want it gone.

The sine wave represents the variation in sunrise or sunset time throughout the year. (I don't think it is quite a sine wave, but close enough). The step function represents the offset from standard time of your clock.

The closer you are to the equator, the smaller the amplitude of the sine wave.

A permanent standard time system is saying "we should approximate the sine wave with a constant".

A permanent DST system is saying the same thing, except with a different constant.

A standard/DST switch-twice-a-year system is saying that we should use a step function that has one value for part of the year and a different value for the rest of the year.

The more steps you use the better you can approximate the sine wave (and since human circadian rhythms sync to sunrise, aligning our schedules to the sine waves is good for us biologically).

But the more steps you use the more hassle it is to deal with the times when the step function changes (you get a short disruption of your sleep schedule and the hassle of changing your clocks--although nowadays most clocks can change themselves).

Thus...you have a tradeoff to make. So far, we've only done one-valued and two-valued step functions.

An interesting possibility is to go to a much larger number of steps, such as 12 or 52 or even 365. With a large number of steps, the change from step to step would be small so no sleep schedule disruptions. With most clocks nowadays having processors they should be able to handle this.

The main problem with this is that it would effectively be getting rid of the 86400 second day. 9:00 one day would not be exactly 86400 seconds from 9:00 the next day. Every time you crossed a step there would be a deviation.

You've got places like Kabul in Afghanistan where the timezone is 1/2 hour off the normal hourly zone (UTC +4:30). You've got countries that want to be aligned with their neighbors for commerce so normal timezone boundaries have exceptions.