And I think it's where the problem lays. What you erroneously call a landslide is a "realigning election" (indeed 1980 was). But a landslide isn't necessarily a realigning election, it just means a very big win.
Though I think this is completely subjective, I think I agree most with this sentiment. And using that definition, I think a landslide is any presidential election where a candidate wins by about 8 points or more. When a candidate wins by that sort of a margin, they're going to take at least 350 or more electoral college votes, more often than not. Plus, to me, it just looks right on paper. 53-47%, with a 305-230 electoral college result just doesn't
look like a landslide to me. But take for example Clinton in '96, where he defeated Dole 49.2-40.7, with 379-159 electoral votes. In my personal opinion, that looks like a landslide to me. As well as 1952, 1956, 1964, 1972, 1980, 1984, 1988, and 1996 - only considering races post-FDR.
I view 1988 and 2008 as the cut-offs. 1988 gets called a landslide because Bush won by +7.72 points, but most importantly, he somehow managed a 426 electoral college rut, to Dukakis' 111. That's pretty significant to me. Obama's pop vote was just slightly smaller than Bush's, +7.27 points, while Obama's electoral college margin was a bit smaller, 365-173.
Again, completely subjective, that's just how I see it.