I think this is fun (and I'm glad to see no longer attributed to Shaw). A lot of people - especially native English speakers who apply these sorts of rules without ever thinking about them - would benefit from thinking a bit more carefully about orthography and how it signals pronunciation correctly in English. Trivial example is double consonant used to signal short preceding vowel (buggle* vs bugle - even though buggle is not a word we all know how to pronounce it).
Made me think for some reason of Freeman Dyson's interesting essay How We Know about information theory and encodings which was posted on HN a few days ago: http://www.nybooks.com/articles/archives/2011/mar/10/how-we-...