Hacker News Reader

Comment by SushiHippie | Hacker News Reader

IMO learning the helix keybindings is definitely worth it. Install it, check with `hx --health` what LSPs you need and install these and that's it. And you still have so much configuration you could do, but don't really need.

View on HN

There is still no plugin support for Helix though right? That's what keeps bringing me back to VSCode, there is always something missing in each editor I look at. Same for VSCode though to be fair, it's sometimes slow on large files and especially when doing type checking, but it's a trade-off I'm making for its DX, I suppose.

I like that Helix's approach is different, and I think the lack of plugins is essential for it to remain on its chosen path. If Helix got plugins, then soon enough everybody would be scratching all their itches using plugins, there would be no users for the core functionality, and it would be the same kind of ecosystem we already have for emacs and Neovim.

What would be ideal for me, but probably not possible in reality, is an editor that cannibalizes popular plugins. So, if a feature is sorely lacking, the plugins appear, but eventually the core functionality grows to replace the plugins. A new user installs the editor and gets a rich set of features that are battle-tested and built to work together. What seems to happen in practice with emacs and Neovim is that the core never gets any bigger, so a new user gets zero benefit from those decades of work, experimentation, and discovery, until they invest the time to integrate dozens of disparate packages into their own setup which they will have to maintain themselves over time.

But that seems to happen in practice with emacs and vim, why would the newly designed editors follow the same rotten path, after all they've managed to break some significant conventions of those other editors?

The other "in practice" side is that without plugins the editor will never get as good as a properly (though painfully, though less painfully with some preconfigured distribution) configured alternative one with plugins

There is a working prototype (that's gone as far as implementing a vim-dirvish picker): https://github.com/mattwparas/helix/tree/mwp-steel-integrati...

The maintainers are collaborating with the dev and hope to get a PR open sometime this upcoming year.

Curious what plugins would you want? The only thing I personally need are LSPs, so it's not a big problem for me.

I'd probably use Helix if it had 1:1 Kakoune keybindings. Or even Vim bindings.

Vim bindings are not the most consistent, but they are ubiquitous. Every program that offers Vim mode has very similar keymap. If modal text editor deviates from them, it better be for good reason.

Kakoune bindings are very different from Vim, but they are provably and objectively [1] better, so that's fine. They are also more consistent and there is a clear idea behind the whole design. It's written down in documentation. You might prefer Vim or Emacs, but at least you can see that changes from well known Vim scheme are not made at whim.

Helix keymap feels like it was improvised without any thought behind it. „Let's take Kakoune binds and add back visual mode cuz I feel like it.” Currently, they are designed by committee in this GitHub issue[2]. I don't see any design notes and explanations why should I spend time learning Helix keymap.

[1]: https://github.com/mawww/golf [2]: https://github.com/helix-editor/helix/issues/165

To be fair, before I used helix, I only used neovims basic keybindings (hjkl, dd, ...) and nothing advanced, so I didn't really need to 'relearn' much, but I completeley understand when someone used a set of keybindings for some time it's really hard to switch.

Reply on news.ycombinator.com