Yes, though that's a little more logical, since perldata specifies that numbers can be written in scientific notation in a form like 12.34E-10. So 0E0 is just a number, and "0E0" is a string that has length (is true) but evaluates as 0E0, which is zero (false). If you want to think about it the other way, '0' is a special case of a string that has length but is treated as false.

What I find interesting about "0 but true" is that it more clearly highlights to a reader that something unusual is going on than something like "0.0" or "0E0" -- and to such an extent that it was excepted from the usual warnings so as to make it easier to use that way. Anyone know the history of this? Or what version of Perl it first appeared in? (timtoady, are you out there?)