RFC3066 bis: use of local codes

Doug Ewell dewell at adelphia.net
Sat Dec 6 06:44:58 CET 2003

Peter Constable <petercon at microsoft dot com> wrote:

> Similarly, re 15924 it says, " ISO 15924 reserves the codes Qaaa-Qtzz
> for private use values. These codes should be used for non-registered
> script values."

The most recent I had heard was that the private-use range in ISO 15924
was to be limited to the range Qaaa-Qacz.  It would be interesting to
know which is correct.

I hope the decision is made to restore the Qaaa-Qtzz range.  This would
constitute 13,520 possible codes out of a theoretical total of 456,976,
or less than 3 percent of the available code space (although of course
we know that some "four-letter words" will never become ISO 15924
codes).  By comparison, the range Qaaa-Qacz is only 78 codes, or less
than 0.02 percent of the available code space.

Restricting private-use codes to the smaller range seriously affects the
potential of these codes to have any mnemonic value, an otherwise strong
point of the four-letter ISO 15924 codes.  See, for instance, my page on
suggested codes for artificial and constructed scripts at:


and note that not only can none of the 37 "Q-codes" bear any real
resemblance to the script name, but that this modest list consumes
almost half the available codes.

It's hard to imagine that every possible use for script codes has
already been considered, such that the larger range can be rejected as
unnecessary.  It's equally hard to imagine that the range from Qada
though Qtzz will be in such high demand by "real" scripts (all starting
with the letter Q) that reserving them for private use will prove to be

-Doug Ewell
 Fullerton, California

More information about the Ietf-languages mailing list