Re: Urgent call for clarification of Armenian numbering rules

From: Kent Karlsson (kent.karlsson14@comhem.se)
Date: Thu Jan 29 2009 - 08:47:27 CST

  • Next message: Julian Bradfield: "Re: Urgent call for clarification of Armenian numbering rules"

    Hi!

    See also http://de.wikipedia.org/wiki/Armenische_Zahlen, which gives
    characters also for 10000 and 20000.

    I don't know anything about Armenian numbers beyond what you have pointed
    to, or can find by web search.

    However, the text in http://www.w3.org/TR/css3-lists/#lower-armenian is a
    bit worrying. It says:

    "The two characters for the 7000 digit should be combined and rendered as
    one character. When the U+0302 character combines with the 7000 characters,
    it does so as if the two characters were one."

    But U-0302 does not, and should not, work that way. If 7000 is really
    represented by two characters that are ligated that way, expecting ALSO that
    U+0302 to work like described in the CSS3 document (not using 1DCD;COMBINING
    DOUBLE CIRCUMFLEX ABOVE and ZWJ), then there is a need to allocate two new
    ARMENIAN LIGATURE characters (upper- and lowercase) for this combination
    rendered as a must-be ligature. But since it seems like that the "two
    characters" for 7000 in Armenian numerals seems to be a mistake (given at
    least two Wikipedia articles that appear to be independently written), this
    seems to be an error in CSS3 rather than anything else.

    http://www.easycalculation.com/funny/numerals/armenian.php also gives just a
    single character for 7000 (though that page does not seem to be entirely
    independent...)

            /kent k

    Den 2009-01-29 11.09, skrev "Richard Ishida" <ishida@w3.org>:

    >
    > CSS 2.1 allows you to number list items using Armenian numbers, but doesn't
    > provide any details about how that works. The CSS3 Lists module does provide
    > detailed implementation advice, but a couple of things have been called into
    > question about how it should work. IE8 CR1 implements numbering exactly as per
    > the CSS3 spec, but other major browsers exhibit differences. We have a very
    > small window in which to check whether the spec is correct before IE8 is
    > frozen.
    >
    > See the description of Armenian numbering in CSS3 [1].
    >
    > See a slightly different description in Wikipedia [2].
    >
    > There are two questions arising from comparing the links above:
    >
    > 1. is 7000 expressed as Ւ U+0552 or ՈՒ U+0548 U+0552 ?
    >
    > 2. is an appropriate default for 'armenian' upper- or lowercase?
    >
    > 3. is the glyph that indicates values over 9,999 a circumflex or a line
    >
    >
    > For results of tests [4] on recent browser versions see our results page [3].
    >
    > Summary [
    > All user agents tested support list-style-type: armenian as far as 9,999,
    > except IE7. However, Firefox and Opera use Ւ not ՈՒ for 7,000, and all but IE8
    > CR1 default to uppercase rather than the specified lowercase. IE8 is the only
    > user agent completely compliant with the CSS3 Lists module.
    >
    > Only Safari, Chrome and IE8 support numbers above 9,999 (although this is a
    > high number for a list, so the impact of that is probably less than for lower
    > numbers).
    > ]
    >
    > If you have knowledge of Armenian numbering, please send your thoughts on the
    > questions above asap.
    > Thank you.
    >
    > RI
    >
    >
    > [1] http://www.w3.org/TR/css3-lists/#lower-armenian
    >
    > [2] http://en.wikipedia.org/wiki/Armenian_numerals
    >
    > [3]
    > http://www.w3.org/International/tests/results/results-list-style-type-armenian
    >
    > [4] http://www.w3.org/International/tests/css/test-list-style-type-2
    >
    > ============
    > Richard Ishida
    > Internationalization Lead
    > W3C (World Wide Web Consortium)
    >
    > http://www.w3.org/International/
    > http://rishida.net/
    >
    >
    >
    >
    >



    This archive was generated by hypermail 2.1.5 : Thu Jan 29 2009 - 08:52:31 CST