On Sat, 11 Jan 2014 22:40:20 -0500, Daniel Bowman <[log in to unmask]> wrote:

>That depends on what diphthongs SL allows.  If it only allows specific
>ones, then you can define digraphs for vowels that will never occur
>together.  This is akin to how English assigns "sh" for  ʃ, since the
>combination /sh/ does not exist in English phonology.  

But it does exist crossing syllable boundaries, _exhale mishap crosshatch monkshood_ etc.  Anyway, I'd disagree with your "since"; although the relative disambiguity of "sh" compared to the old English solution "sc" (and other Middle English contenders like "ss") may well have helped it along, I'd hesitate to call it the reason for that spelling.  The replacements for "sc" first turn up in the early Middle English period when most of the revisions to the orthography can be characterised as Frenchification.  Now French of the time AIUI didn't have any /S/, but it did have /ts tS s/ with a frequent spelling of each being _c ch s_, so that /S/ _sh_ was the natural way for Old Francophones to complete the analogy, again AIUI.  

>This is the strategy
>I used for Angosey.  I needed a way to represent ɯ, but I had already used
>the letters "aeiou" for other vowels.  So I defined ɯ as the digraph "eu"
>in Angosey, since the sound represented by "u" never occurs after the sound
>represented by "e."  So any time I see "eu" I know it is ɯ.  Now, part of
>the angst of my other recent post is that I came up with some other digraph
>combinations pretty early on (10 years ago), and later "discovered" that
>the diphthongs actually DID exist, leading to ambiguities.

I wouldn't mind seeing your list of 38 rules, for what it's worth.  (How e.g. does it compare to Rosenfelder's effort for English, ?)