Both were minor annoyances for me at first lol, thankfully the widget button can be removed (Settings, Home Screen, Edit button - toggle off)
The favorites box is an interesting one, I was originally puzzled with it being empty until I started opening and pinning apps. Now settled on using it as an overflow for my home screen, where 5 of the most used are shown under the clock (with music permanently pinned), and the rest sit in the favorites box.
I did notice some stutter on the apps menu which only occurs while the keyboard is open, but goes away completely when the keyboard disappears. For others it seems like the entire app drawer is stuttery (noticed a long discussion on GitHub with some potential solutions: https://github.com/MM2-0/Kvaesitso/issues/257 )
Free google play credit, I usually get an email every year for it
But I do pay for Plex, despite Jellyfin being a thing. If I like something and it’s worth it to me personally, why not 🤷♂️… but you will never find me defending their kinda crappy decisions like the new Discover feature, removal of “All Songs” from the plex apps in favor of moving people to Plexamp, removing the Gallery sync a few years ago etc.
Some people want their software to be 100% FOSS all-eyes-on-the-codebase, others just do a balancing act based on their personal values.
I value my software to be “transparent enough” in how it operates, “just work”, and hackable to some extent - if I really wanted to I can swap out the ffmpeg binary that Plex uses for transcoding to something else (doesn’t remove the Plex Pass limitation for those curious), I can hook into the server API to change ambient lighting colour based on the cover/background of whatever media is playing, I can create speakers running a Linux board to cast Plex media to, etc. But once that hackable ship sails, then I will look to FOSS alternatives.
For Niagara, everything “just worked”. No noticeable bugs, fast search, consistent feel and design, useful contextual info (e.g. next calendar event shows under the clock), and gestures that made sense for its overall UX. Using it felt less like you were using a “launcher”. The yearly sub was cheap enough that I wouldn’t mind covering for it if I didn’t get credits, and having a single person working on software usually comes with a high level of attention to detail (particularly in performance and UX) but it does have the downside that the experience may be more opinionated and closed compared to if it was a community-driven FOSS project instead IMO.
Alas, google didn’t send credits this year, Niagara made less sense for value/worth-it compared to Kvaesitso, so I abandoned it.
For me, Kvaesitso does everything in a slightly different, much more customizable way, and being FOSS was one of the things that made it particularly attractive as a replacement