Re: Standaridized variation sequences for the Desert alphabet?

From: Michael Everson <>
Date: Wed, 22 Mar 2017 21:33:39 +0000

On 22 Mar 2017, at 20:26, James Kass <> wrote:
> Michael Everson wrote,
>> The old EW and OI and the new EW and OI are clearly *different* letters.
> "Different" versus "variant”?

Yes, different. All of them share the SHORT I [Ιͺ] stroke but the base characters are 𐐅 𐐉 (1855) and 𐐋 𐐃 (1859).

> Michael's analysis seems correct. If Deseret was not already in the Standard, a new proposal for its encoding including eight characters covering the two dipthongs would not be amiss, would it?

Capital and small 𐐦 π‘Ž 𐐧 𐑏 are already encoded. If the other four are required, nothing prevents them from being proposed and added.

> An alternative would be to use the ZWJ mechanism to indicate a preference for the desired letters.

Joining what? We encoded 𐐦 π‘Ž 𐐧 𐑏 explicitly, not as ligatures, though they are in origin ligatures.

> My opinion that variation selectors would be the right approach was based upon concerns about existing data getting "broken". But, if there isn't any existing data…

If 𐐦 is in origin a ligature of 𐐆𐐉 and the 1859 one is in origin a ligature of 𐐆𐐃 then the 1855 and 1859 letters are **NOT** β€œvariants” of one another. They are *different* letters in origin, regardless of their intended use.

The choice to use 1855 EW or 1859 EW is a matter of *spelling*, not glyph substitution. If the later letters are really required, they should be added to the standard. We should not abandon the good precedent we have for character identification just for expedience. That’d be a way to turn the UCS into a glyph registry. :-(

Michael Everson
Received on Wed Mar 22 2017 - 16:34:21 CDT

This archive was generated by hypermail 2.2.0 : Wed Mar 22 2017 - 16:34:21 CDT