Re: Surrogate points

From: Asmus Freytag (asmusf@ix.netcom.com)
Date: Sat Feb 05 2005 - 12:39:41 CST

  • Next message: Asmus Freytag: "Re: Surrogate points"

    At 01:14 PM 2/4/2005, John Hudson wrote:
    >Hans Aberg wrote:
    >
    >>If the
    >>description I got was right, that Unicode tries to enforce that Arabic
    >>numbers should not be represented in the order they are written and read,
    >>but in reverse...
    >
    >It might be a good idea to confirm whether the description is right
    >*before* making public criticisms. I'm not aware of anything in Unicode
    >that enforces what you claim.
    >
    >If you are talking specifically about conventions in Arabic mathematics,
    >as distinct from typical layout of Arabic numerals in text, then one must
    >assume some level of layout above and beyond what is normal for Arabic
    >text, just as one must assume a math handler for European mathematics.
    >When Unicode characters make their way into such a system, the maths
    >handler usually -- and necessarily -- takes control of virtually all
    >aspects of layout, because the conventions involved are so far removed
    >from normal text layout. At that point, whatever Unicode specifies beyond
    >the character code is moot.

    Another and possibly better way to state this is that whenever Unicode
    makes statements about the default layout behavior of characters, it is in
    an attempt to ensure that users and implementers face no ambiguities as to
    which characters or character sequences to use for what purpose. For
    bidirectional languages this includes in which order characters are encoded
    in order to correspond to a specific reading order when displayed.

    A mathematical layout system extends this concept by having (many)
    additional, domain-specific rules, but again, a well designed system leaves
    no ambiguities as to which character codes are to be used for a given
    aspect of the mathematical expression being encoded.

    In other words, the purpose of the Unicode character properties is not
    about enforcing specific aesthetic layout for each scripts, but to enforce
    uniformity in ecoding, so that implementers of layout systems can rely on
    the fact that user would have chosen a particular character sequence for a
    given purpose, and that users can rely on the fact that implementers will
    create a layout for a particular character sequence that matches the intent
    of the user to within the capabilities or technical limitations of the
    implementation.

    A./



    This archive was generated by hypermail 2.1.5 : Sat Feb 05 2005 - 12:42:53 CST