blog: using vim again

This commit is contained in:
Thomas 2024-02-25 11:55:29 +00:00
parent 989b9ea197
commit cd4887510f

View file

@ -0,0 +1,64 @@
---
title: 'Using Vim again, or: how the tool shapes the work.'
date: 2024-02-25T11:54:47.344Z
slug: 2024-02-25-using-vim-again-or-how-the-tool-shapes-the-work
author: Thomas Wilson
---
I don't know, man, every ninth moon cycle something comes over me and asks
"remember how you learned software engineering using vim and rails and that was
sort of it?"
That time has come, all aboard Steam Train Vim.
Specifically, I found [Lazy Vim](https://www.lazyvim.org/), a neovim setup,
which I am configuring with [lua](https://www.lua.org/), a scripting language
I've never used before.
I am already learning stuff !
It's _wild_ how much I learned about making software without Intellisense (or
language servers), integrated test runners, refactoring tools, copilot
autocomplete.
I wrote a lot of my doctorate in plaintext files in vim.
I did a lot of my foundational thinking and building work as an engineer here,
in vim.
But I feel weird that the layer between "me" and the thing that runs my
software is quite chunky.
And the software I interracted with became more and more like the big and
chunky codebases that these tools excel at. I became a professional software
engineer, I guess.
Alongside, maybe because of, this - the act of writing software has become
something chunky. It has become a proper-noun Activity.
I _love_ the JetBrains suite of tools. I do not want to stop using them. They
are so good.
But the spirit of rebellion, of _actually_ hacking away on a computer became
something I don't really do any more.
I am familiar enough with the commandline that it's not magic. I hand-wavy know
what's happening. But I think there's a feedback loop here: I only see things
done with bash, sed, awk (and friends) when either i) it is a trivial example,
or ii) a "hacky but it works" solution that no one's touched since 2012.
The promise of Vim from every Vim-huckster was that you could _edit code at the
speed of thought_. It's an appealing idea, but most of the time I am not
limited by typing speed, I'm limited by thinking speed (and quality).
Problems are just hard sometimes. And [for over a century
now](https://quoteinvestigator.com/2017/02/09/efficient/), we've known there's
no point in optimising for:
> systemic ways of doing things which need not be done at all.
Anyway, maybe I'll do something with it (cool), or maybe in seven days I
realise present/past Wilson is/was being too idealistic and not pragmatic
enough.
Whatever, I'm writing this in Vim and I'm having a great time.