Anyone considering to switch should keep in mind that these are projects with very different philosophies.
Neovim is always chasing the shiny new things; while that's exiting it comes with breaking changes, general instability and the possibility of changes that you might not like.
Vim is the exact opposite. You can drop in a .vimrc from 20 years ago it will most likely work fine. It should be noted that this focus on stability does not necessarily constrain innovation (eg. vim9script), it just sets a conservative pace of improvements.
Lsp support is not shiny and new. I’ve used vim for ages and it’s a huge breath of fresh air to use something not stuck 20 years behind. If you’re the kind of guy who turns JavaScript off to browse the web you’re probably someone who will like old antiquated vim. I of course love vim, original or neo, to the fullest. But I'm also saying it was a massive step up in performance, features, and making vim actually behave like an IDE.
Yes. I tried Neovim last year. Spent 1 day, and did not achieve parity with my Vim setup so I abandoned it. A contributing factor may have been: getting distracted with analogous plugins on the nvim side. Now that Bram is gone, I'll take another day or two and switch over sometime this year. I don't have to though. Vim works. It always has. It probably always will.
The first reason is I’m actually satisfied with vim’s part of what neovim is, so there’s no reason.
But I tried neovim recently. It had worse support for windows (qt-nvim, iirc). It couldn’t rerender or re-column-count itself on resize and also had few other cosmetic issues that I’ve noted but forgot already. One stupidest thing was you can’t explorer-associate it easily in windows because of how parameters get passed between qt-nvim and a host editor process. I regedit’ed it to a working state after a while, but man.
Otherwise I’d try to side-grade to it, but these issues are complete blockers and it’s a little concerning that they weren’t addressed for so many releases.
Maybe something got fixed since, but I only care enough to check once a year or two.
I would add that inertia works at both a personal and institutional level. For example, Vim, Emacs, and Nano are available on most supercomputers but Neovim really is not (haven't seen it myself). Many people may like having highly customized setups --- nothing wrong with that in my view --- but you can't use what your system does not have available. Plain old Vim is still much more universal in many areas.
I have attempted to use neovim, but due to its bleeding-edge development cycle that they often break things, I could not use it even though I have tried really hard.
I have decided to give it time to mature more and become more stable and maybe in the near future I will think about it, at least that's what I said.
Now with the passing of Bram, I have an extra reason to stick with Vim. I plan to study its code so I can start participating in QA-ing the code and help maintaining it in any way I can.
That's the least I can do to show my appreciation and continue his legacy.
Even though I've slowly migrated over to neovim throughout the past ~2 years or so, I still use vim as more of a "general purpose" text editor.
Neovim (mainly due to the amount of useful plugins + lsp support) is used by me as a full featured IDE, which I almost exclusively use for larger coding projects.
For quick scripting or editing of files I use my vim config, which is less bloated and very minimalistic.
The differentation probably doesn't make much sense, but for some reason I like using both.
That experience is very limited, and the word “most” in that sentence is a.most certainly very inaccurate.
Some people are noisily turning vim into vscode. The rest of us love vim exactly for what it is, but love neovim a little more for any number of small or large reasons.
I stick with Vim, as it tends to be pre-installed on many Linux systems in the cloud (e.g. Digital Ocean droplets). I get the same on the server as I use on my MBP locally.
I use Vim because I found my config for Python development seems to work better on both Windows and Linux, without differences between machines. The former is a job requirement. This is probably because I had a Vim setup that worked, but it needed some work to convert to Neovim that I did not have a reason to invest in.
There are some interesting plugins in Lua, but I don’t need Vim to be an IDE. I use VSCode when that is necessary, I don’t need Visual Studio, etc.
Vim works just fine, so I have not seriously (yet) considered switching. Not saying this would never changed, just that I have literally not considered switching (yet).
Still use vim. I use windows at work so old vim is better. Not sure if it's still true today but vim has better windows support. I remember neovim developers celebrating all the windows code they deleted. Congrats on deleting windows platform support I guess. :-|
Have used vim with the same vimrc for c. 20 years, probably will continue for the next 20. What I like about vim is that it just works, and any features I don't need are turned off. My configuration is minimalist, only about 20 lines.
Neovim is always chasing the shiny new things; while that's exiting it comes with breaking changes, general instability and the possibility of changes that you might not like.
Vim is the exact opposite. You can drop in a .vimrc from 20 years ago it will most likely work fine. It should be noted that this focus on stability does not necessarily constrain innovation (eg. vim9script), it just sets a conservative pace of improvements.