Re: Exemplar Characters

From: Antoine Leca (Antoine10646@leca-marti.org)
Date: Thu Nov 17 2005 - 05:04:35 CST

  • Next message: Michael Everson: "Re: ISO 15924: Different Arabic scripts?"

    On Wenesday, November 16th, 2005 23:48Z, Mark Davis wrote:
    >
    > http://www.unicode.org/cldr/data_formats.html

    Re-reading the whole text showed me we might lack something: a textual data
    (I'd call it "00Meta") which could be used to keep the memory of the
    decisions done about the progressive build of a particular locale.

    Then other translators than the first one could refer to it in order to keep
    consistent rules through the whole locale; and the (often acrymonic)
    discussions about style could then be confined in the edition of _this_
    00Meta data, rather than scattered between various threads, where it is
    probably difficult to collect them later.

    Since this 00Meta data is not for final user consumption, it can be updated
    more often without too much harm; on the other side, having it clear would
    help to speed up the updates of the other datas.

    Some examples of the data content (of course, it is expected the content
    will be written in the native language of the locale, I just used English
    here to make my point clear):

    - dictionnaries used as reference (in order):
      - xxx, Somewhere 19xx, xxx pages;
      - yyy, Elsewhere 200y, yyy pages;
      - <URL:http://blahblahblah....> as of 200z-zz-zz

    - we note accents on all letters including capitals!

    - the ^ accent is dropped (according to blah blah blah); the ring is kept on
    words of Scandivian origin [or the reverse, it is just an example ;-)]

    - "Óles" should be written [... in some uniform way!]

    - initial letter is in small case unless it is always written in upper case
    [or initial letter for xxx, yyy and zzz are upper case]

    - please [don't] follow the recommendations of the Whoever reform of 19zz

    - we strictly follow standard SSSsss:19ss, and then we track newer datas
    using Whatever Source; ask SomeContact if youy believe there is a spelling
    error in the standard, don't report a bug here

    - we do not use standard NNNnnnn:19nn, since it is not used in practice, but
    rather blah blah

    - datas for foo and bar have been approved for use by SomeAdministration
    <URL:http://official publications gov...>, care should be exercised if they
    need to be modifed [or] speak with SomeCharge if it needs an update

    - datas for baz and oof have been blindly extracted from the databases of
    AnotherBigBrother, Inc. (which we thank here for sharing it), and should
    probably be reviewed and enhanced

    - character \ughjk (U+ghjk) is intented to be replaced with ? if it does not
    exist in the target encoding scheme

    [for a region-less locale, "qq":]
    - we favour the usage of SomePlace since it is also often used in
    AnotherPlace; the spellings which differ for BigBrotherPlace will go in the
    locale for that region, qq_BB

    - we thank She, He, and She for their reviewing work.

    - SomeAgency greatly enhance the quality of this locale repository by
    allowing He to work blah blah blah

    Etc.

    Antoine



    This archive was generated by hypermail 2.1.5 : Thu Nov 17 2005 - 05:06:12 CST