It depends a bit on the use case. I try to follow naming conventions within specific environments like Python. When just sorting some documents together, I usually do a mix of Kebab and snake case, where I split semantic parts with underscores and connect words with dashes like

2024-08-30_author_document-name_other-important-info.ext

Snake case.

• Starts with a lowercase, good for shell autocompletion
• No spaces, so no worrying about spaces in shell commands
• ‘_’ is better than ‘-’ because it shows the spaces between words more clearly

I put an unnecessary amount of spaces in all my file names to break anyone who wants to use CLI tools on them

i use windows btw

COOLFI~1.AME

Underscore to delineate different parts, hypen to delineate words.

Like: my-resume_draft.pdf

And to make it consistent and easier to reuse parts for project names and such, I have a command line utility written for it. It caches the parts and uses a template system (support for generating current datetime in parts)

Available here (is in AUR too):

https://github.com/Atreyagaurav/nameit

SCREAMING_SNAKE_CASE just has the best name

I like Camel Case for code, but mostly because it’s ingrained in my brain, coming from Java as my first language.

For folders and files, I like Kebab Case.

I am a fan of Python’s or Rust’s official conventions.

For package names, tho, I don’t get why this-is-used over this_clearly_better_system, as I would expect a double click to select_the_whole_thing, whereas it does-not-happen-here.

For files, kebab case. For variables, snake case. For servers, megaman villains.

Camel case, but with a twist – if the next word is about to start after a capital letter, I’ll have it lower case.

topSecretFBIfile.txt for example

IT’S

COOLFILE.NAM

THERE IS ONLY 8.3 AND THERE IS ONLY UPPERCASE

I like to use my enterprise number and a UUID (all in lower case, for legibility). Here’s an example:

.1.3.6.1.4.1.33230.0d456e46-67e6-11ef-9c92-7b175b3ab1f1

Now you might say that the UUID is already globally unique or at least pretty unlikely to turn up anywhere else, so why bother prefixing it with more stuff? To that I say: “I need to be absolutely or at least reasonably sure … OK nearly sure”.

Anyway, you maintain a database of these things and then attach documentation and meaning to them. An editor could abstract and hide that away.

I started this post as a joke. Not sure anymore. Why get your knickers in a twist with naming conventions for variables and constants. Programming is already a whopping layer of abstraction from what the logic gates are up to, another one wont hurt!

How about “cool file name”?

All my systems use modern file systems that are case sensitive and can contain any character except / and \0.

For files? An unholy amalgam of snake, kebab and dot, depending on what exactly I’m trying to convey. (I still have much-ported files around with DOS 8.3 filenames, so they’re truncated scream case (SCREAMIN.NAM), but I don’t actively name files like that anymore.) The important thing is to separate the words/sections while using characters that are valid without escapes in both ext4 and vfat if at all possible.

For variable names, camel or Pascal case (depending on language convention) if I think anyone else is going to read it. Flat case for code I don’t think anyone else is ever going to see (don’t do this—it has left teethmarks in my ass from time to time and will do the same to you).

25% Camel case, 25% Pascal and 50% of the time flatcase. It drives me insane when I try and autocomplete a folder only to realise it was Downloads instead of downloads. I keep telling myself i will go through and make it all flatcase but I put it off because i tell myself i will rebuild my computer next week every week.