I assume the appeal to LXQt isn't so much that it's "lightweight" as that it's "simple". Because, like you said, Plasma is actually very lightweight and snappy, and I'd be willing to bet there are very few places where LXQt will run well and Plasma won't.
Not in my experience, though I have pretty high/less realistic standards coming from i3.
It's plenty usable on modern hardware (anything midrange+ from the past 10 years or so), but it is worth noting that I've still run across folks running core 2 duos in recent times. (Although fwiw they were running plasma, just complaining about it not being very responsive)
If your acquaintances have machines that struggle to run Plasma, I'm honestly surprised they can run graphical apps written in recent GTK or Qt versions at all.
In many cases, Plasma's going to run better/smoother than these "lightweight" window managers and desktops. Because, all the Stockholm Syndrome-ing about "simplicity" often translates to "probably is single threaded and does more CPU work than it should because it's easier." In my experience, they'll tend to have worse vsync issues and tearing.
I doubt i3 is actually drawing shit on the screen faster than Plasma with animations off.
Thankfully I've been able to avoid that for the most part, but I will say it's certainly easier to avoid with plasma's compositor.
>I doubt i3 is actually drawing shit on the screen faster than Plasma with animations off.
I'd definitely be interested in a raw speed comparison if there was a good method/benchmark, but I don't think that has ever been my limitation (although if I were running on lower spec hardware everything counts)
Plasma certainly has more features, but it's still definitely heavier than i3 (or most (all?) tilingWMs), LXQt, icewm, and the like.
On decent hardware there's not a huge point other than preference, but when you've got 2GB of RAM and a 12 year old CPU every bit helps https://i.imgur.com/vZszQFE.png
I could I suppose point a camera or a phone at the screen and measure I suppose.
I'd love to find some good testing on it somewhere, but I do recall off the top of my head that Windows XP or Windows Vista/7 (with Areo disabled) had significantly lower latency than the GPU accelerated counterparts in Vista+. I believe it has to do a lot with Vsync, but imagine there's some additional overhead involved as well.
Of course many would say removing screen tearing is a valid tradeoff there, I mostly agree but not everything is significantly effected by tearing either.