Transliteration, perfection?

Colin, thanks for pointing out the obvious re accents on the small upper-case vowels.
Nice point about the dot over "i" - didn't know that.

What we are coming to, gradually, is a visually-obvious distinction between short and long,
(marked by simple vowel for short, circumflex for long with shewa as raised e but only if vocal)
with the a machine-obvious distinction to mark vav or yod vowels (using small upper-case).

This results in every form of vowels being machine-distinguishable,
except for the distinction between shewa+vowel and the vowel itself (which is unimportant).
And we manage this using only one form of accent is used, which is relatively compatible with other systems. Great!

For tsade I am still torn between ts and tz.
I can see that ts fits better with systems which use s but I prefer tz for pronunciation.
I think of Nazareth - a name which we only knew in Greek (it only occurs in the Gospels),
and everyone thought it was spelt with a zain till archaeologists turned up
a signpost in the 1980s spelling it nun-stade-resh-tau.
Suddenly "he shall be called a Nazarene" made sense
because it wasn't linked to Nazarite vows but to notzer  - the branch of Isaiah,
and the curse against the Notzerim in the Eighteen Benedictions. 

From what you say about diphthongs, I conclude that we can ignore athna etc.

I think have arrived at a solution which fits the best of all worlds:
* displays relatively as relatively standard
* readable
* survives copy+paste
* easily typed. 

See the document TransliterationScheme.doc in the STEP-PROG Dropbox
Also, I've updated the page at http://www.tyndalearchive.com/STEP/TEST/testtrans.htm
to show how the transliteration could be displayed.

David IB