Re: Re: Security Issues: áNavajo

From: Asmus Freytag (
Date: Mon Mar 28 2005 - 00:39:39 CST

  • Next message: Doug Ewell: "Apostrophe (was: Re: Security Issues: ┬áNavajo)"

    At 01:33 PM 3/27/2005, Philippe VERDY wrote:
    > >
    > > U+0027 : 2019 is preferred for apostrophe.
    >Fine suggestion if a text will be typesetted, because it will enforce all
    >quotes to adopt the apostrophe glyph when they are (most often) true

    What the Unicode Standard is saying is that you should use U+2019 when you
    want to encode an apostrophe, such as the one used in English, which is
    typographically indistinguishable from the closing single quote in English.

    It would certainly be incorrect to use that character in any circumstance
    where an incompatible glyph shape is required. So I think this part of the
    discussion is just
    the usual misunderstanding that sadly characterizes so many of the recent
    discussions on this list.

    At 01:57 PM 3/27/2005, Jon Hanna wrote:
    > > So what? Are you suggesting that the French keyboards should
    > > be updated to generate a true apostrophe when striking the [4
    > > ' { ] key, so that most texts will now contain the correct
    > > character?
    >No, because there are contexts where U+0027, U+2018, U+2032 or other
    >characters should be produced.
    >What we need is some sort of machine that can manipulate characters encoded
    >as numbers and do some sort of computation to produce other characters.
    >Given the need for computation in this process I suggest we call such
    >machines "computers".

    An excellent suggestion, Jon!

    Seriously, the "one character - one keystroke" model has been implicitly
    and explicitly repudiated by Unicode from the beginning. The creators of
    the Standard were quite aware that the number of keys is limited, so they
    always considered that users would rely on input methods and similar
    support to help them enter text.

    Software support such as automatic substitutions during text entry can be
    context sensitive in ways that mere keyboard maps cannot. Such support is
    still far from perfect, but it is far superior than fixed mapping, because
    it is easier for users to override the default where not appropriate, or
    even to customize the results.

    There's still ongoing research, for example in the area of mathematical
    text input, which requires rapid selection of a large number of symbols.
    Compared to that, the problem of typing an apostrophe are trivial.


    This archive was generated by hypermail 2.1.5 : Mon Mar 28 2005 - 00:40:54 CST