Re: Script variants and compatibility equivalence, was: Response to Everson Phoenician and why June 7?

From: Peter Kirk (
Date: Mon Jun 07 2004 - 05:44:19 CDT

  • Next message: Peter Kirk: "Re: Revised Phoenician proposal"

    On 06/06/2004 14:38, Patrick Durusau wrote:

    > ...
    > The reason I pointed out that Semitic scholars had reached their view
    > long prior to Unicode was to point out that they were not following
    > the character/glyph model of the Unicode standard.

    I don't claim that they are following the Unicode model. But their
    insight into character identity remains valid.

    > In other words, if you ask a Semitic scholar a question about
    > representation of Phoenician, you are most likely getting an answer
    > based on a criteria other than the character/glyph model of the
    > Unicode standard.
    > That in no way makes the Semitic scholar's answer wrong, in fact is it
    > right, for their domain. It has no relevance at all for a proposal to
    > encode a script based on the Unicode character/glyph model.

    I agree with you that it is the Semitic scholars' domain to judge
    whether the Semitic abjads share the same characters or not - although I
    accept that generalists' and Indo-Europeanists' views on the issue
    should also be taken into account. But I do not agree that this is of no
    relevance to the Unicode proposal. As noted in section 1.6 of TUS, p.7,
    "Before preparing a proposal, sponsors should note in particular the
    distinction between the terms character and glyph as defined in this
    standard." When judging a new character proposal (and a fortiori a
    proposal for a whole new script) the first responsibility of the UTC,
    WG2 etc is to determine whether the proposed new character is in fact a
    separate abstract character or merely a glyph variant of an existing
    character, according to the Unicode character/glyph model. The best
    criterion for making this judgment is how the script is understood by
    the scholars in whose domain it lies.

    I accept that there is a separate argument about the utility of
    sometimes being able to make plain text distinctions between glyph
    variants of the same abstract characters. This seems to be why variation
    selectors and Mathematical Alphanumeric Symbols were introduced. And I
    accept that there is utility in allowing a plain text distinction
    between Phoenician and Hebrew. But if the Unicode standard is to remain
    coherent, arguments from the character/glyph model and arguments from
    utility must be carefully distinguished.

    I do not mean to imply that the current proposer has not noted "the
    distinction between the terms character and glyph as defined in this
    standard." Dr Kaufman is wrong in suggesting that he does not understand
    glyphs or Unicode. As it seems to me, the proposer has rather rejected
    the considered opinions of Semitic scholars that the abjads are made up
    of the same characters, in favour of his own judgment that they are
    separate abstract characters. This is the judgment that I am
    questioning, on the authority of your clear statement, Patrick, of the
    scholarly view that the abjads share the same characters.

    On 05/06/2004 23:59, Asmus Freytag wrote:

    > ...
    > Unlike you, I see no urgent need for Unicode to invent new mechanisms.
    > Implementers will be glad to know that UTC is not contemplating new
    > mechanisms either -- new mechanisms tend to have a way of impacting,
    > by their mere presence, even implementations that don't want to
    > support them, while new characters don't.

    I agree with Asmus here. New mechanisms are undesirable. My point is
    more that there is a need for something to be done, and if the UTC
    rejects as too untidy use of the existing mechanisms to do so, maybe
    they need to consider new mechanisms.

    Peter Kirk (personal) (work)

    This archive was generated by hypermail 2.1.5 : Mon Jun 07 2004 - 05:46:08 CDT