There's nothing obvious about how to use persistent nested variables, or how to assign key/value pairs to a blessed file handle typeglob, or the difference between calling a package subroutine as either a method or a function. The more complicated your data structure gets, the more convoluted ways you can reference it. The subtle yet different ways to do the same thing (or almost do the same thing), like int@_ vs $#_ . Does the camel book explain the magic of $VERSION? Sure it's documented, but you'd have to read every perl doc to finally get to what it is and how it works, and then somehow remember its quirks amongst the pages and pages of docs you just read.

Evaluating true and false, or just simple comparisons, can be a mystery. Heck, just figuring out what's going to be a syntax error or not depending on if a variable was declared, defined, interpolated. Or referencing a nested data structure when one of the elements doesn't exist. Even just general conventions and best practices fills its own O'Reilly book. We haven't even begun talking about regular expressions.... Then there's the autoloader, and constants, namespaces, scalar/list context, signal handling, ....

Part of the difficulty of Perl is that you can write code that works perfectly until the clearly-obvious hole you left is triggered and a fatal error occurs, and then you get to learn about the debugger. It takes years to use all the parts of Perl that have weird magic or complex, non-obvious functionality tied to them. Saying you could use it all without problems after only a few weeks is really misleading.

The fact that there is a "strict" mode at all speaks to how difficult it can be to get the language right.