[omegalulw]

Why? It's an excellent principle, especially for debugging. The biggest benefits of your commits and commit messages are not you but future code reviewers. Imagine someone reads the blame on a file and sees WiP commits (that absolutely will eventually happen), that's nightmare fuel.

[WorldMaker]

As someone who has had to do some "deep archeology debugging" of legacy code, I'd much rather see that git blame full of WIP commits from junior developers than squash/rebase commits from the same. Questions like "was this an intentional regression or bad test behavior from an in progress commit or an accidental regression from a bad merge?" are much easier to answer when you've got "more parts of how the sausage was made" to work with. When I'm not doing "deep archeology" I have tools like git praise --first-parent to focus on the high level (completed PR merge commits).