One problem is measuring what we mean by "paper". I can quite believe this is we include all the fake journals, which feature huge numbers of junk papers. By t

One problem is measuring what we mean by "paper". I can quite believe this is we include all the fake journals, which feature huge numbers of junk papers. By the same argument, 99% of webpages are never visited by anyone but bots, but that isn't a "crisis of the internet", just a lot of junk. It would be nice (and I don't know how) to provide a clearer division between junk and real journals -- I now basically ignore all new journals, unless I know someone involved personally (either as editor, or author). That's the only way I can stay sane, but it doesn't seem like a good system.