Re: ASCII and Unicode lifespan

From: Nick Nicholas (
Date: Sun May 22 2005 - 07:34:22 CDT

  • Next message: Doug Ewell: "Re: Two new Public Review Issues"

    from Michael Everson:
    > At 02:00 +1000 2005-05-22, Nick Nicholas wrote:
    > >The Phaistos Disk (or board game: see .sig), as limited a corpus as
    > >it is and as utterly undeciphered, is not much further along than an
    > >abecedary for all we know, and hardly makes inline glyphs
    > >impractical.
    > Unless you want to interchange the data reliably. ISO/IEC 10646 is
    > intended to support data interchange. PHAISTOS SIGN TATTOOED HEAD is
    > what it is, and all sorts of materials may cite it.
    Same holds for Voynich, or for the roadsigns used in Austria. Just
    because signs belong to a small repertoire of symbols doesn't mean
    they are text, or (*even if they are text*) that they should be
    encoded as such: all sorts of materials may cite a "no left turn"
    sign too. Can data on Austrian road signs be exchanged reliably,
    through some negotiated higher protocol? (Hint: XML entities.) If so,
    then this is no objection to treating Phaistos the same way.
    > >A script used exclusively in meta-discussion of the script, and not
    > >to transmit text in the script itself as a message, doesn't sound
    > >like my idea of plaintext, any more than would a discussion of
    > >horticulture with illustrations require us to encode rose cultivar
    > >illustrations as codepoints.
    > We have lots of characters which don't meet your notion of
    > plaintext, then.
    Yup. And it's not like similar retorts haven't been raised with
    regard to those characters either. (In fact, in my recent proposal
    for epigraphical Greek letters to Unicode, I raised such a retort
    myself :-)
    > Nonsense. This is not a dismissal of the character-glyph model. But
    > that model is just that -- a model. It guides us. It does not bind
    > us. Otherwise all sorts of legacy cruft would not have been allowed
    > to be encoded. And that cruft is still being added to the standard,
    > in big great buckets.
    Which is why I said in the next paragraph "ab initio". When we don't
    have legacy cruft to deal with, why should the character-glyph not be
    followed rigidly? If it's just a guide, why not make special
    allowance for that Serbian T, even if no existing charset
    differentiates it from Russian T? That's slippery slope stuff.

    And as is clear from
    n2096.pdf , the bodies reviewing ab initio proposals such as
    Hieroglyphic can be rather rigid about the character-glyph model. For
    obvious reasons.
    > >Where Unicode is considering encodings ab initio, with no such
    > >cultural or legacy static, it should take its own rules seriously.
    > >The Phaistos Disk (or board game) is not the Han character
    > >repertoire.
    > No, but neither is it irrelevant to European culture and literature.
    Michael, I'm from Crete myself; and how culturally important the
    artefact is to me or to anyone else does not alter my opinion that it
    is not ready for encoding as a script.
    > >After all, just because we got saddled with oodles of precomposed
    > >codepoints through legacy doesn't mean we should dismiss the
    > >avoidance of new precomposed codepoints for being "rigid and
    > >rule-bound"; the case looks to me fully analogous.
    > Not at all. A precomposed letter with diacritical marks is disallowed
    > because it can already be represented by a string of characters.
    > Phaistos characters cannot be represented by a base letter and a
    > diacritic.
    Not my point: Unicode has a rule on unique representability, which is
    not violated for cruft, and which is followed rigidly. The politics
    and the resources are not there to police the glyph-character model
    to the same extent for Han. But those conditions do not apply here,
    and there is no good reason why a conservative approach to the glyph-
    character model shouldn't rule Phaistos out of bounds, given that a
    new Phaistos disc could cast everything upside down tomorrow. If the
    PUA was good enough for Klingon, after all, it's good enough for
    > I think this shows that you haven't really looked at the Phaistos
    > disk corpus. Small as it is, I think it's pretty clear that none of
    > those characters is a glyph variant of any other one.
    Medial and initial Arabic meem. And to the inevitable retort that
    inspection of an undeciphered Arabic would make it obvious that
    characters are only medial or final: the disc is a corpus of 60 words.
    > You're playing
    > devil's advocate, I suppose.
    Sure. As will DIN. Then again, Phaistos was defeated once already, so
    there's little point really devil-advocating one way or the other.

    It appears to be a real script (or a board game), and there are
    people who want to be able to work with the script as part of the
    decipherment process.  On the other hand, there *is* just the one
    document (or board game), so there's only so much one can do.
    (John Jenkins on the Phaistos Disk; Unicode mailing list)
    Dr Nick Nicholas.

    This archive was generated by hypermail 2.1.5 : Sun May 22 2005 - 07:35:17 CDT