Re: Directionality Standard

From: Otto Stolz (
Date: Thu Jan 10 2008 - 11:17:40 CST

  • Next message: Waleed Oransa: "RE: Directionality Standard"

    Hello Murray Sargent,

    Jony Rosenne observes:
    > See Unicode Standard Annex #9, The Bidirectional Algorithm, clause 3.3.1.
    > "The Paragraph Level".

    You have written:
    > This is the context rule I mentioned. It doesn't handle paragraphs that start
    > with a strong character of the opposite direction from the desired paragraph
    > direction.

    In these cases, the author should insert a strong character of the
    desired direction into their plain-text sources, at the beginning of the para-
    graphs in question; U+200E and U+200F are precicely made for this purpose,
    cf. <>,
    and <>.

    If a higher level protocol is used, you’d rather exploit its features.
    E. g., in HTML, use the DIR attribute,
    cf. <>.

    In any case, the author has to express his intent.

    Note also that the author has to insert control characters
    to tell the bidi algorithm the border between the inented
    runs of opposite directionalities, where a sequence of
    punctuation marks (or other characters having no stronhg
    inherent directionality) spans a region of opposite
    directionalities. A common example being a quote of the
    other direction: The quotation marks belong to the sur-
    rounding text, but the final punctuation of the embedded
    quote belongs to the latter, hence needs opposite directionality.
    Cf. <>
    for an example (even an amusing one, if you can read
    German, and possibly Yiddish.)

    The only shortcoming I can see, is that the Bidi algorithm
    for plain text applies to only one paragraph. Hence, to specify
    RTL direction for a whole chapter or document, the author
    must either use a higher level protocol, or take the pains to
    mark the direction of every single paragraph that happens to
    start with the wrong character. But this is the price to
    be payd for setting one’s heart on plain text, I reckon ;-)

    Best wishes,
       Otto Stolz

    This archive was generated by hypermail 2.1.5 : Thu Jan 10 2008 - 11:19:59 CST