I wish there was more Computers PhD+beyond research that was like, "I am going to explore how to make zsh more comprehensible and devote 3-5 years to this" for open source projects that need better documentation or rewrites-of-the-project-but-dont-currently-have-any-plans-but-everyone-thinks-about-it.
all of it is basically technical writing, techno-sociology, UX design, user studies, research, and hands-on testing -- and universities could be doing this but instead we keep sending them towards high-level math instead of doing these community service projects to pay for their sin, cos, and tans