Re: Common Locale Data Repository Project

From: Mark Davis (
Date: Fri Apr 23 2004 - 21:48:36 EDT

  • Next message: John Cowan: "Re: Common Locale Data Repository Project"

    You can reiterate it all you want; in practice, 3066 tags are used as locale
    identifiers. And for a narrow sense of locales, that is perfectly reasonable.
    For a broad sense of "locale", including timezone, user's currency, religious
    preference, etc., it clearly would not be reasonable, and I would agree with you
    for that.

    >ISO 639 is not unstable. It is an open code set that is being added to
    over time, but I don't think that should be referred to as unstable --
    that term suggests other things.

    ISO 3066 has *demonstrated* instability, because they remove codes, then reuse
    those codes for different entities. It'd be like our removing a character, then
    later putting a different character in that spot*.

    ISO 639 has not yet *demonstrated* instability. They have removed codes, but
    since they haven't reused them, one can handle that with an alias table, keeping
    all the old codes usable. However, there is no policy documented *anywhere* that
    says they won't. As long as they don't have that, and given the demonstrated
    instability in ISO 3066, the standard simply cannot be trusted to be stable in
    the future.

    * Yes, I know we did that for Korean, when we were first getting started. But we
    learned from that, and put into place firm policies against that ever happening
    in the future. We have no such assurances from ISO, for some pretty key
    components: language codes, country codes, currency codes, or script codes.

    ► शिष्यादिच्छेत्पराजयम् ◄

    ----- Original Message -----
    From: "Peter Constable" <>
    To: "Mark Davis" <>; "Philippe Verdy" <>;
    "Unicode List" <>
    Sent: Fri, 2004 Apr 23 16:18
    Subject: RE: Common Locale Data Repository Project

    > From: []
    On Behalf
    > Of Mark Davis

    > You are talking about Locale IDs. There is currently work underway on
    an RFC to
    > replace 3066

    But let me reiterate from my correction to Philippe: even the
    replacement of RFC 3066 is a specification for *language*
    identification, not *locale* identification.


    Peter Constable
    Globalization Infrastructure and Font Technologies
    Microsoft Windows Division

    This archive was generated by hypermail 2.1.5 : Fri Apr 23 2004 - 22:20:18 EDT