There already is a corollary to TMTOWTDI in this sense; it's the First Virtue of Programming. Laziness means expending a little energy at the outset to save a lot of work later on. I think the camel book does a good job discussing how laziness can be useful, I think it even discusses CPAN in that section. But "get thee to CPAN" isn't going to help in this case.
Here's a challenge: using CPAN, try to find an indication that posix.pm is the right module to solve the problem leons did with his subroutine, without specifying strftime in the search box. I couldn't do it. The closest I came on CPAN was a reference to Date::Calc, on a search for 'date' and 'sprintf'.
If I cheated and used a google search restricted to perldoc.com, the same query gives me Date::Calc as the first choice, and a Tk:: module that mentions strftime as the second choice (if I'm willing to dig for it).
I suspect it is often less work to reinvent the wheel.
To reframe the question: how can we make CPAN more helpful to lazy people?
Here's one possible solution: CPAN should contain a copy of all module documentation in a format that google will index-- all the pod files for every module. Right now I think a lot of the documentation is locked up in .gz files.
Perhaps then, when I do a web search for "perl date sprintf" CPAN will show up in the first 40 results. I suspect that would go a long way toward making people aware of CPAN.