I leave a lot of TODO and FIXME comments in code (which are then tracked and highlighted by my IDE). These range from code that may actually be broken on certain rare or unexpected corner cases (though not dangerously so), to known inefficiencies, to areas where the rough idea for a much better approach exists but can't yet be fully developed.
Very roughly, they are a measure of known technical debt.
Same here (in gvim). I also like pylint's default behavior of issuing warnings where it finds these and penalizing the code's final score in the report -- it's a proxy for code problems that have been silenced or worked around, but still need some attention later. Integrating this info with a bug tracker seems like a good idea too for sufficiently complex projects.
I think I wind up adding new ones at about the same rate as old ones disappear, but I'd have to run a report over the history to know for sure.
It's always nice when I come across a note that's been made irrelevant by other parallel improvements along the same theme -- even though the note wasn't the impetus for the other changes. It's like finding a $20 on the street!
Very roughly, they are a measure of known technical debt.