An entry on acme's use.perl journal linked me to his String-Koremutake distribution, and the docs for the String::Koremutake module in turn pointed to this document on shorl.com, presumably the source of the specification.
What interested me most was the Q/A pair "But it all looks like gibberish to me. How am I supposed to remember any of it unless I am a native swede or something?".
I can only imagine that the authors are not native speakers of English, or at least that they speak a rather different variety of English than the ones I'm most familiar with.
For starters, I find their example of "O as in mOon" strange; I would have expected <O> to represent an [o]-ish sound, not a [u]-ish one (which is what "moon" has for me). (Maybe a Swede would think along different lines; IIRC, <o> is roughly [u] for them?) On the other hand, they do say that MO is to be pronounced as in moose, so who knows.
But next, their key word for U is "sUe"—which is exactly the same (phonemic) vowel for me as "moon"! (John Wells's lexical set GOOSE.) At least, if we ignore the yod that I have in "sue", but I hope that's not what the pronunciation key was supposed to get at. Maybe it was?
But what really struck me was that they list S as being pronounced as in "Silly", but SY as in easy—but "silly" has /s/ for me while "easy" has /z/!
Also, "E as in Eternity" is ambiguous for me, since depending on stress, that word starts for me with either schwa or /i/. While the dictionaries cited on dictionary.com s.v. eternity use the KIT vowel for the first vowel of that word—yet DE is said to be pronounced as in destination, which has DRESS in the first syllable for me!
All in all, it seems like a rather odd pronunciation guide.
If the authors were, indeed, not native speakers of English, then I would think: that's not a problem per se, but perhaps they shouldn't be making pronunciation guides if their grasp of English phonology is, shall we say, not that well aligned with what I think it is.
(Or is there a joke I've been missing?)