5 ms·
1. Install Spacemacs. 2. Observe a flurry of errors every time you start emacs. Ignore and hope it works because no idea how to fix any of it. 3. Find, among
by clarry 6y ago
1. Install Spacemacs.
2. Observe a flurry of errors every time you start emacs. Ignore and hope it works because no idea how to fix any of it.
3. Find, among a million others, the command that seems to do what exactly what you want. Command gives an error. Dunno how to debug.
4. Watch emacs features that should work ootb not work because spacemacs or evil or one of the three bazillion packages did something that broke it. Dunno how to debug.
5. Complain on HN, get told you need to be using the master/develop version of packages because upstream devs don't care about maintaining stable releases.
6. Use unstable packages. Observe new errors whenever you update or do anything. Hope it doesn't break more than it's already broken.
- pantulis 6y agoThis. While Spacemacs is an admirable effort, I think it misses the point: Emacs is an editor for power users from another era, so it requires huge effort by the user to become truly productive by todays standards. Trying to create layers upon layers of "friendlines" and eye candy is not going to solve the complexity. I've always said that it's better to grasp Emacs with a vanilla config or a small starter kit and see if you can make it grow step by step in a journey that takes years. vim is a similar rabbit hole, maybe less deep. The real question is: can VS Code make you productive enough in less time?
- throwanem 6y agoPretty much, yeah, from what I've seen. Against the shockingly low ambient baseline of "editors that aren't Vim or Emacs", VS Code's highly usable OOTB experience and broad extension ecosystem are enough of an improvement to constitute a real win despite being so poor by comparison with a thoroughly evolved configuration of either of the Two True Editors. I stick with Emacs because, now having invested more than a decade in learning to get the most out of it, I can do a lot of things in seconds that take VS Code users minutes or hours. But the converse is also true, and I think will only become more so. Especially in the realm of remote collaboration and mentorship, which has never been more important than it is today, VS Code does things that Emacs simply can't, and almost certainly never will. That's fine, of course. That different tools should specialize in different things is perfectly reasonable, and I don't have the kind of emotional attachment to Emacs that would give me cause to be upset with its dwindling user share or its lack of broad appeal. It serves me very well, but it's not something I advocate, although of course I mentor those who take an interest. I am looking at picking up VS Code, with suitably Emacs-esque keybindings, for the mentoring-in-programming aspect of my role. That's where pretty much everyone else is, or is going. And in that context, its less fluent and less extensible user interface is probably a boon, once I get over being frustrated by it. In the context of teaching someone less experienced, acting with the speed of thought is really best avoided; if you don't explain why you're doing what you're doing, or offer the chance for questions and discussion, essentially the entire point is lost.
- cdrini 6y ago> I can do a lot of things in seconds that take VS Code users minutes or hours. Out of curiosity could you provide some examples of some of these?
- jamespo 6y ago"Out of curiosity" = "I'm going to nitpick this to hell and back"
- cdrini 6y agoAhh, that's not what I was going for. I'm actually just curious. I'm willing to give emacs a try, but it just seems like it has kind of a steep learning curve, so want some strong reasons to know if it's worthwhile. This was the best I could find googling: https://www.reddit.com/r/emacs/comments/8h1cxa/any_long_time_emacs_users_tried_vscode/ https://www.reddit.com/r/emacs/comments/8h1cxa/any_long_time...
- throwanem 6y agoI mean, a lot of it is just the kind of complex one-off text transformations that would normally require a throwaway script or program to express. (They do in Emacs too, but you can write it as inline Lisp forms in the replacement side of a PCRE find-and-replace, and not deal with any boilerplate.) Refactors and stuff like that all happen through LSP providers, just like in VS Code. If you're looking for a strong reason to think Emacs is worthwhile as a totally new user, I'd point instead to Magit, an extremely powerful and comfortable git porcelain, and Org-mode, which is simply the most powerful and flexible single-user notetaking/outlining/live code notebook tool yet created. Those are the two Emacs features I see most often mentioned as the subject of sentences like "I don't use Emacs for anything else, but I do use it for X because nothing else comes close". That said, as I mentioned above, I'm not really here to evangelize Emacs. It definitely does have a steep learning curve, enough so that unless you see a clear killer feature (in my case, TRAMP's transparent remote file editing, since I edited remote files a lot in those days), it's not likely to be worth the trouble.
- emidln 6y ago> I've always said that it's better to grasp Emacs with a vanilla config or a small starter kit and see if you can make it grow step by step in a journey that takes years. I moved to Spacemacs becaues I wanted to write clojure without maintaining my own Clojure IDE in vim+tmux (this was around 2014, maybe 2015). I was productive in Spacemacs in around 10 minutes. I went from never using Emacs to fixing bugs and writing extensions in Emacs in less than two weeks because Spacemacs pervasively uses which-key to provide discoverability and had vim-style shortcuts I could use for real work (rather than learning a new set of keybindings). The groupings of things into layers makes studying what ecosystem packages provide what functionality much easier than blindly googling "how to do X in emacs" and getting 3000 "unique" (read: half-baked if you're lucky) solutions. This idea that you should suffer for hours/days/weeks building your init.el from scratch while adapting to terminology, keybindings, and a lack of familiarity with the ecosystem is absolutely insane. If you can use vi, you're better off with Spacemacs.
- ryukafalz 6y agoAs a counterexample to that somewhat... I initially tried learning emacs with spacemacs, and the amount of customization in spacemacs’s config made it difficult for me to compare it to much of the emacs documentation out there. Whatever documentation spacemacs had at the time was confusing enough to me that it wasn’t quite enough. I later ended up starting with a stock config, and I probably could pick up spacemacs now if I wanted to, but at this point I don’t feel the need.
- brmgb 6y ago> vim is a similar rabbit hole, maybe less deep. Vim was for a long time and still pretty much is mostly a text editor. It is light, work well with large files and is installed on most Unix machines. I don't think trying to turn it into an IDE makes much sense but if you want to develop with it, installing a language server client takes two minutes and you don't need much more. As a text editor, Vim is really nice. Modal editing can be very pleasant if it clicks with you.
- aldanor 6y agoIf you install Spacemacs from develop branch, don't change the defaults and enable just the git plugin, it's pretty much always going to work just fine unless elpa is down or something. I install Spacemacs in every machine for the sole reason of using Magit with vim bindings. It's so far beyond all other git clients it's even hard to explain to others sometimes. If Magit existed as a native standalone app, I would be happy to pay €50-100 for it, right away. But, but... :/
- michaelmrose 6y agoYou can always donate the same funds to the development of magit.