Re: CSUR Tonal

From: Doug Ewell (
Date: Mon Jul 26 2010 - 09:55:30 CDT

  • Next message: Mark Davis ☕: "Re: ? Reasonable to propose stability policy on numeric type = decimal"

    "Luke-Jr" <luke at dashjr dot org> wrote:

    >> These are not really combining marks; they appear to be nothing more
    >> than ordinary Latin superscript letters. As such, I would suggest
    >> not only that the "multiplication" and "division" superscripts be
    >> unified with each other, but that they be unified with
    >> already-encoded Latin superscript letters S, T, b, m, r, s, and t.
    > Correct me if my understand is wrong, but I believe them to be
    > combining because they can and should "attach" to some degree to the
    > unit character preceding (in the case of division) or following (in
    > the case of multiplication) the divisor/multiplier. It is for this
    > font-encoding purpose that I created two sets which can attach in
    > their specific direction.

    I believe your understanding is wrong. The superscripts are indeed
    intended to appear "relative to" the letter symbolizing the base unit,
    but they are not combining characters in the sense that the resulting
    combination is meant to be treated as a single unit. I don't see any
    "attaching" behavior, even in the sense of something like an 'a' with
    diaeresis, where the diacritic isn't physically connected to the base

    > Aren't superscript/subscript letters compatibility-only as well?

    They have a compatibility decomposition to a character with markup.
    That isn't necessarily the same as a character which is sufficiently
    similar to another encoded character that it would not have been encoded
    except to provide 1-to-1 mapping with another character set. U+00B2 is
    an example of the former; it might well have been encoded even if it did
    not appear in countless other character sets. U+3380 is an example of
    the latter; it almost certainly would not have been.

    > Other than the possibly combining needs, I hope an exception (if
    > needed) could be considered so as to not require rich text simply to
    > type up a unit.

    I haven't proposed that any rich text be required. A superscript
    letter, representing the multiplier or divisor, before or after the base
    unit would be plain text.

    >> multiplication and division signs
    > I only have Nystrom's book as my guide on this, and it does not
    > specify how exactly the higher-order multipliers and divisors are
    > supposed to be written abbreviated. It is possible they were not meant
    > to be at all, but this theory would create problems in present day
    > when such numbers are more commonplace.

    There isn't anything to encode here. The logic of prepending or
    appending a small superscripted letter (by whatever means) would
    naturally extend to higher-valued prefixes or suffixes.

    Note that this problem doesn't stop there; the tonal-system mechanism of
    inventing short words for higher orders of multiplication is unspecified
    beyond the (decimal) quadrillions, which is inadequate for many
    scientific needs.

    > How might I phrase it such that it is clear the extra space is only
    > for a later discovery of Nystrom's extending of Tonal?

    Just make them reserved, like any other reserved code points. It
    doesn't seem obvious at all that subsequent undiscovered works by
    Nystrom will fill these perceived gaps.

    Doug Ewell | Thornton, Colorado, USA |
    RFC 5645, 4645, UTN #14 | ietf-languages @ is dot gd slash 2kf0s ­

    This archive was generated by hypermail 2.1.5 : Mon Jul 26 2010 - 10:00:26 CDT