Re: Seemingly duplicated radicals, reasoning?

From: John H. Jenkins (
Date: Wed Dec 26 2007 - 13:15:43 CST

  • Next message: Upshur Whittock: "A weird character I saw last week"

    On Dec 22, 2007, at 7:56 AM, Jeroen Ruigrok van der Werven wrote:

    > What I understand is that the Kangxi radicals are the base 214
    > radicals as
    > they appeared in the Zihui and Kangxi dictionaries.
    > Nowadays, due to simplifications and other changes, various additional
    > radicals got introduced in order to make look ups easier/more
    > consistent/whatever. For this we encoded the CJK Radicals Supplement.

    Actually, no. The encoded radicals have nothing really to do with
    lookup (in the computer sense). They're mostly derived from the
    custom in dictionaries in their radical tables to include the multiple
    shapes a radical may take. We used to do the same in our radical
    stroke charts but stopped with the chart in TUS 5.0 to save space.

    You really, really, really should think of the encoded radicals as
    dingbats. They're only there so that people can print radical
    charts. (Well, that and compatibility with CNS 11643, which was the
    original motive for the 212 radicals they include.)

    As such, minor differences between the shapes can be significant.
    These are dingbats, after all.

    > U+2e95 and U+2f39 - radical snout (two) (a bit dubious one, since
    > the latter
    > seems to have the bottom stroke drawn past the standing stroke)

    That's the extent of it.

    > U+2ed1 and U+2fa7 - radical long (one) (no apparent difference)

    I'll have to check if this was deliberate (I doubt it) and file a
    glyph erratum.

    > U+2ee3 and U+2fbb - radical bone (no apparent difference)

    U+2EE3 is the Simplified Chinese version (box on the left inside the
    top). U+2FBB is the form everybody else uses.

    > U+2ee4 and U+2fc1 - radical ghost (no apparent difference)

    As before, I should double-check this.

    Basically, since these are dingbats, you may adapt their shapes
    slightly as needed. For example, I would expect that U+2ED1 and U+2EE4
    would be drawn with one fewer stroke.

    On the whole, you will almost never need to use *all* of the encoded
    radicals, because not all of them will be relevant to your locale and
    use. Use the ones you need and ignore the others.

    > Since I am going to use radicals for an application I am developing
    > I want to
    > be sure I am not misunderstanding anything. If the above are indeed
    > not
    > different would I just have to make sure I use the Kangxi radical
    > over the CJK
    > Supplemental one?

    Again, if you *need* to express a difference, then it's possible. You
    should never feel obligated to use any of these beasts, let alone all
    of them. They are dingbats. They don't really *mean* anything.

    John H. Jenkins

    This archive was generated by hypermail 2.1.5 : Wed Dec 26 2007 - 13:18:51 CST