Re: New to Unicode

From: Doug Ewell (
Date: Mon Jul 24 2006 - 09:17:45 CDT

  • Next message: Curtis Clark: "Re: Proposal to encode an EXTERNAL LINK symbol in the BMP"

    Don Osborn <dzo at bisharat dot net> wrote:

    > So which is operant, the IANA registry or ISO 15924
    > or both? Somehow I
    > got the impression that IANA codes were more or less incorporated in
    > the latter.

    For identifying a script in the general case, you can use ISO 15924 on
    its own. For the specific case of a language tag that contains a script
    component, such as "ko-Latn", you should use the IANA Language Subtag

    The relationship between ISO 15924 and the script subtags in the
    registry is more stable than for other ISO standards, mostly because ISO
    15924/RA is committed to stability, but also because scripts are
    inherently a stable concept (at least compared to countries) and because
    15924 is still so new.

    Stephane Bortzmeyer <bortzmeyer at nic dot fr> wrote:

    > If I understand well, the IETF decided to create a separate registry,
    > hosted at IANA, because ISO is not stable enough. (The Internet
    > requires stable identifiers, something that the ISO bureaucrats cannot
    > grasp.)

    In its partial defense, ISO 3166/MA is primarily concerned with keeping
    up with the current state of affairs, not stability. This causes
    problems for applications that do require stability.

    There are at least 3 main reasons for putting all of the valid subtags
    in a single registry:

    1. For stability, as Stephane mentioned. When an ISO Maintenance
    Agency withdraws a code element, the corresponding subtag does not
    magically disappear from the IANA registry; it stays there forever,
    possibly with its status changed to Deprecated.

    2. To give users a single, readily accessible source for all subtags
    valid under RFC 3066bis. Having to go to several different sites run by
    ISO MAs is bad enough, but there are also unofficial lists of ISO codes
    that are not always accurate or up-to-date. This was an intentional
    design goal.

    3. To ensure that the subtags would be freely available and not
    constrained by the distribution policies of ISO or the various MAs.

    > An happy consequence is that the registry is available online (which
    > is not the case with most ISO standards.)

    For the core standards that were used to generate the registry -- ISO
    639, 3166, and 15924, and UN M.49 -- the basic code lists are available
    on the Internet, but the standard itself is not. Even the official
    online code lists don't always provide important additional information.
    Did you know that the ISO 639-2 code elements "qaa" through "qtz" are
    available for private use? How many people know the list of allowable
    private-use ISO 3166 country codes? (Answer below.) This is all laid
    out in the registry.

    > AFAIK, all ISO codes are in the IANA registry, but not always the
    > reverse.

    In the future, if an ISO MA creates a code element that conflicts with
    an existing subtag, it may not be added to the registry.

    Doug Ewell
    Fullerton, California, USA
    Editor, draft-ietf-ltru-initial

    This archive was generated by hypermail 2.1.5 : Mon Jul 24 2006 - 09:23:38 CDT