L2/06-020

Public Review Issue #82: Representation of Gurmukhi Double Vowels

In older Gurmukhi, some texts use two vowel signs on a single consonant; for example, one can find ga with both the oo and u vowel signs. A priori, this can be represented in Unicode using two different sequences:

U+0A17 ਗ GURMUKHI LETTER GA
U+0A4B ◌ੋ GURMUKHI VOWEL SIGN OO
U+0A41 ◌ੁ GURMUKHI VOWEL SIGN U

or

U+0A17 ਗ GURMUKHI LETTER GA
U+0A41 ◌ੁ GURMUKHI VOWEL SIGN U
U+0A4B ◌ੋ GURMUKHI VOWEL SIGN OO

Because one sign is above the base and the other is below the base, both sequences have the same rendering. In general, this situation is not a problem, as combining marks above and below have non-zero, distinct combining classes, and the combining sequences are therefore canonically equivalent. In this particular case, the combining marks have both a combining class of 0, and the sequences are therefore not canonically equivalent. This makes the interchange of text more difficult, and the UTC would like to resolve this situation. Because of the policy on normalization stability, the most obvious approach (change the combining classes of those characters) is not possible. It seems that the best approach is to list a preferred ordering, and the UTC would appreciate input on which sequence is preferable. Please keep in mind that other sequences may present the same problem, may be in other Indic scripts, may be in other non-Indic scripts, so a rationale for the choice that can be used in the other cases, should they arise, is preferable.