I've heard that the standard Lemmy UI is not under active development because the Lemmy UI developers are working on a rewrite. I looked at it for a while though, and I thought that, aside from some missing features and polish, it was fine.
I'm trying Photon right now, and it is also fine, and a little more feature-complete, but it is visibly janky in some respects. Maybe it is my biases, but I also much prefer the model of clicking on things and getting a new page over a single-page app that you interact with via controls.
What was so terrible about the vanilla Lemmy UI? Do people really have a strong level of dislike for it? It seems to me like it just needs some love to smooth down the rough edges and awkward spots, but the core doesn't seem in any way terrible such that it would need to be abandoned.
It's buggy. Submitting a post sometimes takes half a second, and sometimes it takes 30 seconds. Logs show a plethora of errors. The Docker container will eventually crash and restart. It's all part of the Lemmy instance admin experience.
No options to collapse/customize side panels.
UI doesn't refresh after saving settings.
Settings aren't sticky from page to page on occasion.
If I could code I'd pitch in and help. Heck, I'd learn, but the documentation is pretty sparse.
1 I have never seen. The backend was incredibly buggy in 0.18 and early 0.19 versions. Maybe the frontend was too? I've only been running it for a short time but I've never seen it crash yet.
2/3/4 I count as polish things. Yes, they're not ideal. How is a rewrite into a new core supposed to make that better, as opposed to throwing the maturity level back to square 1 and introducing a whole plethora of new little polish things to worry about in addition to those?
I'm not saying you're wrong. Maybe about 1 in the present-day codebase. I definitely see things that could be improved and this is a good list of items, but what I'm asking about was the decision to abandon the codebase and start fresh specifically.