I use TODO in my code all the time as shorthand for: "This code is functional, but if you are going to do another iteration you may want to consider the following improvement or optimization."

It's not at all meant to be to be like an item in a TODO list. Code would be a terrible place to keep that.

The lead programmer at my very first coding job taught me to put a string of at least three hash marks in a comment, to denote “this is a thing which is okay during development but absolutely must be changed/fixed before shipping this product”.

The more essential the change/fix, the more hashes. Made it really easy to do a global search for them, and you could easily filter the results to only show the most important ones by searching for longer strings of hashes.

Twenty years later and I’m still doing that in my personal projects.

At my work we just file a bug and make sure it gets prioritized as a must-fix... The code might contain a bug number and remark too for redundancy, but the bug itself generally spells out the bits of code that need to be changed/removed/whatever.

Using code comments really seems like the wrong place to prioritize changes. Even on personal projects.

I agree in principal, though if the intent is to have the minimal amount of side-tracking while marking something as TODO for the future, I don’t think you can get better than a comment.

A nice cli interface for creating tickets in GitHub/Jira/etc. would help with that. I wonder, would it be nice to have a pre-commit hook that scans for TODO comments, removes them, and drafts tickets? Then both flows would work well.

I suppose it depends on how far you trust your ability to keep everything in your head, and reload everything coming back, and follow through on notes to self. I make todos too, but I try to honor a self-agreement that they be short lived and ideally not make it very far beyond my machine.

I linked https://github.com/dspinellis/git-issue in another comment but it gives a sense of the command line experience. I've also at times hacked up various scripts to create or modify items in github/jira/our internal POS, it can make things more likely to hit the tracker than to be met with a shrug. (It's so odd how sensitive programmers can be to even minor blocks in flow like a few seconds waiting for a webpage, or avoiding refactoring a method / class because in Java a new method might best belong to a new class which to do things 'proper' often entails a new file and then another new file for unit tests, adding them to version control, and all that takes like 30 seconds even with IDE support and is annoying.) I've at times thought about a small vim command that would call one of those scripts to create a new issue pre-templated with the file and location I was on, create, then insert the work item id, but never got around to it. As a pre-commit hook something to look for TODOs explicitly might be interesting, though I generally dislike hooks.