Hey everyone!
I was just wondering why one would use any of the programs I mentioned above instead of VS Code to write Code.
Can someone give me a brief overview of the advantages?
Thanks in advance!
If you frequently work on remote systems you frequently only get command line access, where you can still use vim/nano/emacs but not a full IDE like VS Code. In that case you might find it more convenient to learn one text editor well and forgo the IDE.
you can use vscode over ssh, without installing anything on the remote system.
There’s a long list of caveats when running VS Code over SSH. By comparison, text editors:
- Work on Alpine remotes
- Work on older distributions, and other *NIX systems
- Have no problems with SSH key passphrases or security keys
- Only require a few MB of memory
- When run in tmux are largely resilient to SSH connection issues
Yea, no. It doesn’t work with ssh agent and it cannot read includes and other configuration options. I believe it also tries to install some components remotely which is bad enough, but causes additional issues in environments with proxies or without internet access at all. Iirc also plugins must be installed remotely to work.
In a normal professional context it just does not work and it is a hassle to deal with. It might work in a home lab, but nowhere beyond that.
From the perspective of someone who uses Visual Studio Code, but also knows how to exit vim, there are a couple reasons that most developers who prefer one of the three, at least those I’ve spoken to.
- VS Code is a Microsoft product, and while “open source” it isn’t really open source. The core utility is but Microsoft ships the final application with some proprietary features. If this is your main gripe, then you can try VSCodium instead, which is a “fork” that doesn’t have the Microsoft additions.
- VS Code uses Electron, which is essentially browser emulation and isn’t exactly optimized. CLI editors like the above take up far fewer resources than a Visual Studio Code instance would. Unlike point (1), I don’t think there’s really a way around this in all practicality. It’s just an unavoidable fact. You can chose to still use VS Code of course, most personal computers can easily handle the load. But many see that as unnecessary when they get the same amount of “power” from a CLI editor.
- Plugins for the CLI applications are very powerful, and the ability to navigate using only the keyboard is by design. Many swear by keyboard-only operation of a computer because it’s faster and promotes more optimal methods of doing tasks. It forces discovery of new features and hotkeys by making things annoying to do otherwise. VS Code (and most editors) include a “
vimkeybindings” specifically for this reason. You’ll find that it’s a very popular method of working.
Really it comes down to personal preferences and what you “grew up” using. It’s really hard to transition into something like vim and it takes a concerted effort to switch by most users. You have to want to switch, otherwise you’ll find it too difficult a learning curve or find yourself wandering back to more “featured” applications.
There are likely more reasons out there, but these are, in my experience, the primary reasons.
They work in terminal. To me a more interesting question would be, why VSCode over PyCharm for Python for example.
Vscode with plugins works for many many languages good enough and the git integration is good. So if your job is like mine where you support many systems then having a common tooling is nice. Also you can setup containers with different configurations all ready for various projects.
Vim/neovim with plugins does exactly the same but uses less system resources. Lunarvim is a good place to start for a preconfigured neovim IDE.
Emacs will also be similar but I’m not as familiar with it.
But does actually work like an IDE? I for example love PyCharm understanding type annotation. Not only it highlights errors, but also improves autocompletion and makes big refactoring less scary.
The integration with data grip (unfortunately that’s available in paid version) allows for similar behavior with SQL contained in strings. Which IMO fixes the impedance mismatch that created the need for query builders and ORM frameworks.
VS Code has some pretty good ide features for python, including understanding types, highlighting errors and warnings, linting, navigation features such as go to definition or go to references, and basic refactoring capabilities like rename symbol. These features are enabled by the python language server (pylance, in this case, which is Microsoft’s proprietary one).
You can also get the same features in other editors that support the language server protocol. For example, I use neovim and my setup supports those same IDE features I used to use in VS Code for python.
Because Microsoft is The Debil.
Different tasks. VSCode is littered with half-baked spaghetti code from various projects. My terminal window is for system-level non-project interactions.
I stick with terminals and vi for various day-to-day administrative things, because you never know when you’ll have to log into a different machine which doesn’t have all your software and preferences set up. I would prefer not struggle to recall keyboard shortcuts I last used 100 years ago when that happens.
It’s bad enough having ADHD and long covid to completely obliterate my working memory, without the added irritation of googling “man vi move to section”, knowing you have definitely used the command many times in the past.
