Re: Dutch IJ, again

From: Doug Ewell (
Date: Mon May 26 2003 - 19:49:09 EDT

  • Next message: Philippe Verdy: "Re: IPA Null Consonant"

    Pim Blokland <pblokland at planet dot nl> wrote:

    > My feelings at this moment are that if you do want to make the
    > difference clear, write U+0133 for the "normal" ij sound and i+j for
    > when it's supposed to be two letters, and don't use any tricks such
    > as non-joiners.

    I thought the last time we had this discussion, there was more or less
    agreement among the experts that we should NOT be encouraging the use of
    the U+0132 and U+0133 for new text. They aren't canonical equivalents
    of I+J and i+j, only compatibility equivalents, so not all text
    processes will be expected to match them with the I+J and i+j that
    people have been using to type Dutch for decades.

    Furthermore, they aren't available on most keyboards, so end users might
    find this suggestion a bit difficult to implement.

    U+0132 and U+0133 are present in Unicode because they were present in
    existing character encodings, with which a one-to-one mapping was
    considered desirable. They are not present to encourage ambiguous
    spellings of Dutch words.

    Instead of encouraging the continued (or resurgent) use of U+0132 and
    U+0133, we should teach our text-processing software to titlecase I+J
    and i+j "correctly" in a Dutch context (denoted by whatever means;
    previously I would have suggested Plane 14 language tags). As a second
    priority, software should recognize any existing instances of U+0132 and
    U+0133 as equivalent to I+J and i+j in searching and sorting.
    Recognizing such equivalences is not the same thing as encouraging
    continued use of the compatibility characters.

    -Doug Ewell
     Fullerton, California

    This archive was generated by hypermail 2.1.5 : Mon May 26 2003 - 20:29:53 EDT