Re: Lowercase in one script, capital in another

From: Kenneth Whistler (kenw@sybase.com)
Date: Tue Jul 05 2005 - 18:56:03 CDT

  • Next message: Eric Muller: ""The Languages of China" by Ramsey"

    Chris,

    > >> The Heiltsuk language uses a lowercase Greek lamda ‘λ’ for the [dl]
    > >> sound,
    > >
    > > Well, saying it's a *Greek* lambda presumes the very question you're
    > > leading up to.
    >
    > I say Greek just so people will know what the symbol looks like.
    >
    > > I would propose LATIN SMALL LETTER LAMBDA. We really don't want a case
    > > pair from two different scripts. (BTW, there is a latin small lambda
    > > *with stroke* -- U+019B.)

    This is definitely going to cause problems. U+03BB *is* the Americanist
    lambda, and is recognized as such by the Handbook of the IPA, as
    well, even though lambda (and barred-lambda) are not official IPA
    usage.

    If you need to introduce an uppercase for it for Heiltsuk that is
    *not* U+039B, that will introduce the same kind of confusion that
    is already present in the standard for various overlapping
    case-pairs involving Latin schwas and barred- or hooked-d's. And it will
    introduce further confusion about whether the Greek or the Latin
    lambda should be used in NW Indian orthographies.

    You are also going to run smack into the current firestorm of panic
    regarding visual spoofing security issues. Anything that looks
    remotely like any letter already encoded is headed straight for
    the blacklists, and will end up disallowed in at least some contexts
    that the users might have expected it to be allowed.
     
    > The small lamda with stroke is also used in this language, the λ being
    > voiced [dl] and ƛ being voiceless [tɬ]. There will also need to be a
    > capital for ƛ, but as this is already Latin script, I didn’t mention it.

    Inventing new case pairs for characters that didn't have them before
    is a dicey prospect at best. They will end up being implemented
    badly (and late), I fear.

    > > Mixing scripts is bad. Mixing case pairs between scripts is not allowed.

    There is no such rule. That is an assertion by Michael of existing
    practice in encoding.

    It is unlikely that the UTC would ever assign default case mappings
    across scripts, but that does not prevent someone from using
    an orthographic rule of that sort. Not that I think it would be
    a good idea, but neither Michael nor the UTC nor SC2 defines what
    communities *can* do in devising orthographies.

    --Ken

    >
    > Great. Then I will get back to the communities with this information and
    > start the proposal process.
    >
    > Thank-you
    >
    > Chris



    This archive was generated by hypermail 2.1.5 : Tue Jul 05 2005 - 18:57:27 CDT