[agallant]

I've also seen similar usage of glitch versus bug, and think it's a useful distinction. But I'd say that actually it can be consequential from an engineering perspective, and can even run counter to the outsider intuition - that is, "doesn't work at all" is usually easy to diagnose and often not that hard to fix (e.g. "we just need to update the cert/restart the db/etc."), while "weird every-other-Thursday glitch" is harder to pin down and usually more intertwined with system logic ("why does our videogame physics system cause characters to float but only on levels 3 and 7?").

Basically, "complicated system doesn't work" is often a less intricate problem than "complicated system almost works." The perception may be opposite because it seems like more work to fix something that is "more broken", but that's the same logic as LOC as a good productivity metric. As with all things, there are exceptions, but I'd say the tendency is that glitches are harder to fix than bugs.