Re: [idn] Reality Check

["Adam M. Costello" <amc@cs.berkeley.edu>]

"D. J. Bernstein" <djb@cr.yp.to> wrote:

> Are you suggesting that the WG stop considering any encodings for
> which there isn't at least one letter in the output per letter in the
> input?

A few hours ago the working group stopped considering any encodings
without seven supporters, which I think leaves only DUDE, AMC-ACE-Z,
and maybe MACE (I hope we'll hear an announcement soon).  In each of
these encodings, every input letter is represented by a sequence of
output characters, the last of which is always a letter.  The last
proposal being considered by the ACE design team that didn't have this
property was LACE, which lost support on Jun-05 in favor of DUDE (which
is simpler than LACE, a little less efficient for Han and Hangul, and a
little more efficient for everything else).

It has been discovered that ACEs that use an integral number of output
characters per input character are simpler and need not be less
efficient, and once you're in that regime, it's easy to arrange for the
last character to always be a letter.

AMC