[MonkeyClub]

Same here, I find the KWh to be a very useful unit for daily calculations of that sort, where I want a quick upper bound on things to make estimations.

Last year I switched off my fully-functional 2008 workstation (a lovely Fujitsu Celsius W370 on OpenBSD, a furry joy) because of such an upper bound difference (300W vs 65W for the ThinkCentre that hides among the books on my desk's side).

This sort of works in a similar way with light bulbs as well. Although lumen would be the appropriate unit for luminosity, the packaging uses wattage to indicate luminosity.

Although lumens and Watts are correlated, they aren't dimensionally equivalent as Joules and Watts are (CMIIW).

That "100W" on the package an electrically 14W bulb simply means "it's only using 14W, but shines like a 100W bulb, go ahead, BOGOF".

[andrewla]

> 300W vs 65W

A watt to watt comparison is fine. Why hours? I can tell you right now that the big one uses ~4.5x the power. Is it really that much easier to convert the time you're using the device to seconds? If you're going to multiply by the electricity cost anyway, might as well break out the calculator one step early.

[trelane]

I use watt when I want to compare power, and watt*hour when dealing with energy. Hours because hours are a lot closer to the real spans of time I use, and therefore much easier to calculate with. E.g. it's more common to run a 3kW AC for 3h than 3s. So, 9 kWh instead of 9 kWs (or 180kWs, for the same time span)

> If you're going to multiply by the electricity cost anyway

I don't usually convert to money. If it's a linear cost per kWh, then I can deal with that at the end or, more likely, don't actually really care.

The _actual_ cost you pay is often tiered anyway, so who knows what the price _actually_ is until you have the whole month's worth of power usage finalized. And then it's not clear what I could account to what tier.

> might as well break out the calculator one step early.

Bold to assume I use a calculator much.

[MonkeyClub]

Yeah exactly, I didn't even need to take time into account, I was looking for the upper bound presuming it's always on with a simple Wattage comparison.

But that was mostly because that workstation, albeit lovely, had a very long boot process.

The ThinkCentre boots in under a minute, so I actually end up only booting it when needed.

(Some of my work can be done offline, and I jump at every opportunity to `halt -p` and be in the quiet offline space/state.)