https://www.forbes.com/sites/jasonevangelho/2019/10/23/bold-...
What we get from older technologies is better tuning.
I get it, I should be using FOSS drivers, but I do a lot of work involving neural networks and CUDA, so Nvidia's drivers are non-negotiable.
Unity doesn't get in the way of me interacting with my desktop applications and is fairly utilitarian compared to GNOME and KDE based desktops.
The GNOME project seems to have made a conscious decision to go for the consumer market, which usually cares more about things "just working" than their being tweakable. That's a good thing - that market needs to be served - but it's not what I wanted. A big part of the reason I run Linux is for detailed control of my user experience, and KDE is the better product in that regard for my use case.
> KDE is an utter delight for everyone else
The way you frame it as OS X convert vs everyone else is kind of rude. I would say the hardcore Linux customizer is the minority, and everyone else prefers the easy to use out of the box UIs of OS X and Windows.
To be fair, this probably says more about me than it does about either environment.
But if you try i3 or OoenBox, they are way faster
But those are just window managers no? You can run them alongside KDE or Gnome.
It's questionable how much value is added by running a plasma session over simply running KDE apps and services.
> Canonical's write-up of the work done to improve GNOME is also interesting for the long list of what didn't work and how wrong some of the developers' initial assumptions were—especially regarding GNOME's use of JavaScript. > It turns out JavaScript had next to no effect on performance. So sorry JavaScript haters, but "assuming that JavaScript is slower than everything else written in C" is evidently a bad assumption
As this is really hard to fix they fixed this by constantly running the gc in December 2018. This worked OK but didn't fix the issue for all until 2019.
Does making it broken for 8 years so usability comes 8 years late count as making it slow.