I would humbly suggest that something like “lines of application code ever written”, or better yet, “source code files evaluated since Unix epoch [or “per second” if you prefer]” is probably a better metric. I think in that analysis, you would find that tabs are more than holding their own.