Re: Encoding Personal Use Ideographs

Date: Sun Nov 04 2007 - 20:14:06 CST

  • Next message: Philippe Verdy: "RE: logos, symbols, and ligatures"

    Quoting Andrew West <>:

    > On 04/11/2007, John H. Jenkins <> wrote:
    >> I swore I wasn't going to post anything more to the threads on this
    >> topic, but since James is a rational and reasonable human being, I'll
    >> break my vow.
    > I wasn't either, but since I'm neither as rational nor as reasonable
    > as James is I'll break my vow as well.
    >> 1) IDSs are inadequate to provide acceptable rendering;
    > Not if they are used to render specific [as yet] unencoded characters
    > for which a font has precomposed glyphs.

    Myabe the point here is that there are many characters that cannot
    being encode clearly using IDS, which is something eveyone agrees on.

    IDS are adequate for many unencoded characters as long as the user is
    aware that if later encoded there will be compatibility issues.

    >> 2) Normalization of IDSs would be necessary for operations such as
    >> searching and sorting and is an absolute nightmare; and
    > Yes, but my understanding is that James is not seeking to introduce a
    > general-purpose compositional model for the Han script, but is simply
    > exploring the possibility of using IDS sequences to represent specific
    > [as yet] unencoded characters is a manner that TUS
    > <> says is valid
    > and which learned members of this list have confirmed is acceptable.
    >> 3) It is generally felt that making up *new* ideographs is a Bad Thing
    >> and ought not be encouraged.
    > I agree, but where would the Han script be if no-one ever created a
    > new character or two?

    As the FAQ acknowledges new characters will always be coined, this is
    the same way as new words are added to English. Whilst using a
    different spelling of a word to evryone else is in general a bad
    thing, adding you words is something every living language does.

    >> I won't say that the UTC or WG2 will absolutely oppose any and all
    >> compositional models for Han, *but* the bar is very high indeed. This
    >> approach has been bruited in Unicode pretty much since the beginning,
    >> and the UTC is well aware of the practical problems involved. A
    >> compositional model won't fly in the UTC unless these problems are
    >> addressed; for all practical intents and purposes, the UTC would
    >> require an actual implementation, at the very least.
    > I hope I'm not the only one here who is now totally confused. Why
    > should it be any business whatsoever of the UTC if James as a
    > respected font designer wants to map specific IDS sequences to
    > precomposed glyphs in his fonts or even if he or anyone else decides
    > to create a font that composes IDS sequences into unitary glyphs on
    > the fly ? When the standard already says, in quite some detail, that
    > this is acceptable behaviour why should anyone need to go running cap
    > in hand to the UTC to get their permission to do so ?
    > I'm half tempted to drop everything else I am doing and take up your
    > challenge to produce an "implementation" (whatever that means in this
    > context), despite the fact that we poor Windows users are stymied by a
    > rendering engine (aka Uniscribe) that only allows IDS sequences that
    > comprise IDC characters and CJK radicals, but not IDC characters and
    > CJK unified ideographs and/or CJK radicals as the Unicode standard
    > specifies.

    I guess this makes me a frog at the bottom of a well.

    > Andrew (aka U+2FF4 U+4E95 U+86D9> ???)

    This message sent through Virus Free Email

    This archive was generated by hypermail 2.1.5 : Mon Nov 05 2007 - 09:42:38 CST