Vowel romanization

Tirelat has a high central vowel /i\/ (or /ɨ/ for those with Unicode
compatible email software), which I've been writing as "y" in
romanization (http://www.io.com/~hmiller/lang/Tirelat/script.html has an
old version of the Tirelat script and romanization, which is pretty
close to the current version except that i've been using the circumflex
instead of the acute accent for vowel length). I've considered using the
dieresis to represent central vowels, but this would be problematic if
you want to add tone or vowel length with diacritics. Also, there
wouldn't be any way to distinguish between low-mid and high-mid, but
that's not much of a problem for central vowels.

If I keep "y" as a high unrounded central vowel in Tirelat, I'll have to
come up with something different for [y]. I've considered dotless i, but
that won't work with accent marks (they'd look like regular accented
i's), and would require using the capital dotted I (as in Turkish: İ) as
the regular capital version of /i/. So I'm considering U+0197 (Ɨ) LATIN
CAPITAL LETTER I WITH STROKE, which uses U+026A (ɪ) LATIN LETTER SMALL
CAPITAL I as the lower case form.

Still, it doesn't seem possible to fill every vowel slot with a unique
letter. There are a few obvious possibilities, but there are also some
annoying gaps.

            Front   Central   Back
Close       Ii  Yy   Ɨɪ       Ưư  Uu
            Ɩɩ                    Ʊʊ
Close-mid   Ee  Øø       Ɵɵ   Ơơ  Oo
                     Əə
Open-mid    Ɛɛ  Œœ   Зз           Ɔɔ
            Ææ
Open        Aa

Another problem with these modified letters is that they're practically
impossible to represent in ASCII. One alternative would be to use dots
(and other marks) below letters -- a period could be used to represent a
"dot below" in ASCII, and the central vowels could be represented with
two dots below the letter.

Close       Ii  Yy   I̤i̤  Ṳṳ   Ưư  Uu
            Ịị  Ỵỵ                Ụụ
Close-mid   Ee  Øø   E̤e̤  Ɵɵ   Ơơ  Oo
                     Əə
Open-mid    Ẹẹ  Œœ   Зз  O̤o̤   Ợợ  Ọọ
            Ææ       A̤a̤
Open        Aa  Œ̣œ̣            Ạạ  Ɔ̣ɔ̣

Or I could try to maximize the number of vowels that can be represented
without needing unusual fonts, and not worry about putting tone and
length marks on vowels.

            Front   Central   Back
Close       Ii  Yy   Ïï  Üü   Ưư  Uu
            Ịị  Ỵỵ                Ụụ
Close-mid   Ee  Øø   Ëë  Öö   Ơơ  Oo
                     Əə
Open-mid    Ẹẹ  Œœ   Зз       Ợợ  Ọọ
            Ææ       Ää
Open        Aa                Ạạ  Åå

I don't think I'll need the vowels represented by IPA [ɶ] (CXS [&\]) or
 [ɞ] (CXS [3\]), so this system looks adequate. But that means I'll need
to double vowels for length.

Some examples of Tirelat words with long central vowels:

pýri (pŷri) "brief": pɪ̄ri pī̤ri pïïri
dłý (dłŷ) "ginger": dłɪ̄ dłī̤ dłïï
tǿr "orange": tə̄r təər

Reply