Re: ISO 15924 codes reserved for private use

From: Doug Ewell (
Date: Sun May 23 2004 - 19:45:06 CDT

  • Next message: Doug Ewell: "Re: ISO 15924 codes reserved for private use"

    Philippe Verdy <verdy underscore p at wanadoo dot fr> wrote:

    > Someone said here that there are today lots of more scripts studied
    > than have for now no interchangeability, but that may be still needed
    > for bibliographic references, so that there was already a private
    > registry of private use script codes (PUSC) nearly filling all the
    > PUSC space allocated in ISO 15924 (50 rows, with Code in Qaaa to
    > Qabx).

    I asked about this a year ago, when the private-use space in the draft
    had been shrunk to 78 codes (Qaaa-Qacz) from something much, much
    larger. Now 78 looks roomy by comparison.

    I doubt that there are only 50 entities in the world, including "real"
    scripts, "artificial" scripts, and other things for which people might
    want to use ISO 15924 (e.g. combination codes like Hrkt), that do not
    have an ISO 15924 code. Considering the vast encoding space available
    (456,976 possible codes, minus certain unassignable "four-letter words"
    and other reserved codes), the range Qaaa-Qabx seems unnecessarily
    small, to say the least.

    > I understand that this space is limited to 50 numeric codes, but why
    > isn't there space for 10 times more with 4-letter codes? Is it
    > required that all 4-letter PUSC codes have a corresponding numeric
    > code?

    There are already 456,000 possible alpha-4 codes and only 1000 possible
    numeric codes. So even if there is such a requirement, I don't
    understand why the range of assignable alpha-4 codes should be limited
    to 50. After all, the alpha-4 codes available for normal encoding
    aren't restricted to Aaaa-Abnz (1000 codes) to match the numeric limit
    of 1000.

    > If so, why not reserving all codes starting by "Qa" for private use?
    > This would give 676 rows, whose only 50 first ones have a standard
    > 3-digits numeric code, and could allow librarians to create their own
    > private references by giving them full freedom on the last 2 letters
    > of alphabetic codes, so that they could use them mnemonically (for now
    > all they have is "aa" to "az" and ba" to "bx").

    Just allow any of the Qa codes to be associated with any of the numeric
    codes 900-949, with no prior restrictions. That's the way normal ISO
    15924 codes work.

    Michael, I apologize for raising this on the public list, but nothing in
    our discussion last year (April 6 and 7) convinced me that restricting
    the private-use range to Qaaa-Qacz was necessary. And now it is even

    -Doug Ewell
     Fullerton, California

    This archive was generated by hypermail 2.1.5 : Sun May 23 2004 - 19:47:13 CDT