L2/15-010

Feedback on proposal L2/14-304 to encode BENGALI LETTER AW
Shriramana Sharma, jamadagni-at-gmail-dot-com, India
2015-Jan-26

The proposal is based on the current Unicode practice of encoding atomic 
characters for Indic vowel letters *without* canonical equivalence even when 
visually the character is composed of a sequence of other characters. This 
practice (or at least the absence of canonical equivalence) is an unfortunate 
legacy which cannot be corrected now due to stability policy. As a result I do 
feel that this character is a valid candidate for encoding based on existing 
norms.

However, I feel the following points deserve further patient consideration:

1) The choice of the name as LETTER AW:

There are other North Indic vowel characters (especially in Devanagari) named 
LETTER AW. It is not clear if this is indeed the Bengali script parallel of 
those characters. This should be verified further before the name is allotted.

2) The alternate choice of the name as LETTER SHORT O: 

As per earlier documents submitted by Anshuman Pandey, (L2/09-320, L2/09-377 and 
L2/09-403; see also my L2/10-471) it is known that the DEVANAGARI SHORT E/O was 
an adaptation by Hoernle from the Bengali script. However, the use of the same 
glyph for Kokborok does not mean the character should be named SHORT O. The 
above suggestion of AW seems much better and may be accepted after verification.

3) The choice of the codepoint as 0992 corresponding to DEVANAGARI SHORT O at 
0912:

This is also not really meaningful because the putative correspondence to SHORT 
O is doubtful. The proposed atomic character is better placed at 09D6 next to 
its own vowel sign at 09D7.

4) The suggestion to change the name of 09D7 to VOWEL SIGN SHORT O:

A name change is clearly not possible under stability rules. An informative 
alias may be added to 09D7 with an appropriate name after verification as noted 
above.

5) Sorting order:

Again, the *glyphic* correspondence of the character to the DEVANAGARI SHORT O 
does not mean the sorting order should be the same. The proposal does not 
provide sufficient linguistic information as from dictionaries to determine the 
correct sorting order, especially since the *linguistic* equivalence of the 
proposed character to any existing Devanagari character (whether AW or 
otherwise) is not yet clear. That is what would be helpful to determine an 
appropriate sorting order.

6) Sample text:

The proposal does not indicate exactly *where* in the sample text the relevant 
character occurs. It is not easy for non-natives to determine satisfactorily 
which of the Bengali letters is supposed to be the proposed character. Further, 
no dotted circles are visible in the text, leading one to wonder about the 
claimed absence of support from software for the sequence which is glyphically 
equivalent to the proposed atomic character.

-o-