An experiment with phonetic writing

[eggyk]

Isn't the premise of this that the french person would potentially understand the word once re-spoken by the person transcribing? If we assume that most of the sounds were roughly correctly interpreted by the transcriber, it doesn't matter how they wrote it down, it only matters that they can correctly reproduce the sounds they intended to represent within their system. (the lack of a d in "ujuvi" indicates that this has not taken place)

So if someone were to write "ajawdwi" to mean "aujourd'hui", regardless of what others can interpret by reading it, as long as the writing system consistently reads "ajawdwi" as "aujourd'hui" it doesn't matter. 

Show ajawdwi to a dutch person and it reads "ayavdvi", and perhaps to other english speakers it may be "ahjordwih". It only matters that I -as the person creating the system- can reliably reproduce the sounds that I intended to store.

The same would apply to contractions: "aub" would be read as "au-b" or "oub" to those not in the know, but as long as the person writing knows that it represents "alstublieft"- and reads it as such - the information has been preserved.