On the contrary, I'd say both of these posts are discussing memory use by Perl threads and how it compares to forking. Forking benefits from copy-on-write, while threads do not, meaning that forking tends to use much less memory. This also means that the approach for conserving memory as much as possible is exactly the opposite in each one: load as late as possible for threads and as early as possible for forking.