Re: Telugu vs Kannada confusables

From: Shriramana Sharma (
Date: Mon Nov 29 2010 - 03:46:51 CST

  • Next message: Jeroen Ruigrok van der Werven: "Re: UNICODE version of _T(x) macro"

    On Mon, Nov 29, 2010 at 12:02 AM, Christopher Fynn <> wrote:
    > Do the folded domain names get displayed?
    > If so, doesn't this have the potential to break some rendering
    > systems?

    I don't fully understand you. As I pointed out, the folding is purely
    an internal thing done during the search for duplicate domain names.
    The folding will not change the domain name itself. (Possibly in this
    way it is different from the case folding where all domain names when
    registered fold to lowercase and are registered as a sequence of
    lowercase characters if I am not mistaken.)

    Anyway, this confusable folding as I understand it is much like the
    case folding that is done when you are doing text-insensitive search.
    The text in the input field of the search box is not changed, and
    neither is the text in the document, but it is only that they are
    considered equivalent because the internal processing of the search
    command folds them to a uniform representation.

    So if at all folding produces mixed script text, such mixed script
    text will exist purely as a character stream in the internal
    processing of the registrar's server where the server searches for an
    already existing domain name that folds to the same character stream.
    It will never be displayed. It is only used to determine whether the
    server should respond "Hey there is already a name that is confusable
    with this so I can't let you register this domain name you are asking
    for" or it should go ahead with registering the requested domain name.

    As for the requested domain name itself, it is unlikely to
    intentionally contain a mixed-script text, so long as the requester is
    not intending to create a phishing website spoofing some other
    legitimate domain name! As somebody pointed out already, legitimate
    domain names are expected to have all their characters in the same
    script (but that's a "should" instead of a "must" it seems). Anyhow,
    when you are registering a legitimate domain name, you are unlikely to
    chooes a sequence of characters that leads to bad rendering... The
    server itself will register the domain name (if available and not a
    confusable with a pre-existing name) with the exact characters you
    requested, so that should not be a problem.

    (If anybody more technically knowledgeable than me finds fault with
    any of this, please do correct. I am by profession NOT a domain name
    registrar or web-IT specialist.)


    This archive was generated by hypermail 2.1.5 : Mon Nov 29 2010 - 03:52:27 CST