Re: The Phaistos Disc

From: Kenneth Whistler (kenw@sybase.com)
Date: Wed Apr 05 2006 - 18:52:44 CST

  • Next message: Donald Z. Osborn: "Re: CLDR: Bad exemplar chars for some locales"

    Elliotte Rusty Harold wrote:

    > John H. Jenkins wrote:
    >
    > > I would be perfectly happy with this, BTW. IMHO the disk is unlikely to
    > > be deciphered and I doubt we'll ever see another example, so
    > > functionally the elements can be treated as symbols without full support
    > > for them as a developed writing system. If and when I'm proven wrong,
    > > we can revisit the issue.
    >
    > That's my primary concern. Can we revisit the issue?

    Yes.

    > I don't mind
    > encoding the characters, but it seems extremely unwise to assign them
    > irrevocable categories until we know more about the script.

    There is nothing irrevocable about assigning a character gc=So and then
    later deciding it should be gc=Lo.

    If you want a case to prove the point, consider

    UnicodeData-4.1.0.txt:

    2132;TURNED CAPITAL F;So;0;ON;;;;;N;TURNED F;;;;

    UnicodeData-5.0.0d10.txt (not yet released, but in beta):

    2132;TURNED CAPITAL F;Lu;0;L;;;;;N;TURNED F;;;214E;

    That because it was discovered that this was the uppercase we
    needed for the Claudian letter. It changed in General_Category and
    it got the case mapping to the new lowercase character encoded.
    (And also changed bidi category.)

    There is also nothing revocable about not deciding to give some
    character a specific new value for the Script property. If
    someone proves a true requirement for a Script=Phaistos value,
    nothing is stopping the addition of that at some point in
    the future. I just don't see any clear requirement being
    postulated here.

    There are things which can*not* be changed once decided.
    Among those are code points, character names, and decomposition
    mappings (or lack thereof). But General_Category and Script
    property values are not subject to such absolute stability
    guarantees.

    --Ken

    > Personally I
    > do think we'll one day have more examples of this script. If I were sure
    > that wouldn't happen, I'd say go ahead.



    This archive was generated by hypermail 2.1.5 : Wed Apr 05 2006 - 18:58:35 CST