RE: Arabic encoding model (alas, static!)

From: Kenneth Whistler (kenw@sybase.com)
Date: Wed Jul 06 2005 - 19:11:33 CDT

  • Next message: Gregg Reynolds: "Re: Arabic encoding model (alas, static!)"

    Ashraf,

    > From: "Bob Hallissy"
    >
    > >The idea of a generative encoding for Arabic was discussed and debated by the
    > >UTC over the course of at least a year and a half  (Jan 2002 through to as late
    > > as June 2003). This is the "recent proposal" to which I referred. In the end,
    > > however, this particular proposal was rejected by the
    > >UTC
    >
    > Sorry to ask again: where is the documented rationale for this rejection?

    Asking again isn't going to turn one up.

    The participants on this thread have been presenting the various
    rationales that led to the non-acceptance of the "recent proposal"
    that Bob Hallissy is referring to. That is basically the best
    you are going to get.

    I think you may have some fundamental misapprehensions about how
    standards committees in general operate, as well as the UTC and
    WG2 in particular.

    The presentation of a proposal to a standards committee isn't like
    an application for a license from a government agency (or something
    similar), for which the applicant has some kind of legal entitlement
    and rights to appeal and rights to explicit provision of reasons
    if the application is turned down.

    A proposal in a standards committee initiates a process of
    consideration and discussion, which may or may not result in a
    formal ballotting involving some aspect of the proposal. Failure
    to reach agreement on a proposal (either unanimous or by a voting
    consensus process of some sort) does not obligate the committee
    to then provide a formal rationale why agreement could not be
    reached. It is simply a failure to reach agreement to do whatever
    may have been proposed.

    In the ISO process, once a particular standard or change to a
    standard is in formal ballotting, national bodies may provide
    technical or editorial comments accompanying their ballots. The
    editor of the relevant standard provides formal written disposition
    of comments, and those dispositions generally do provide rationales
    for decisions taken on those comments -- as agreed by the relevant
    working group. You can find many such examples in the ISO JTC1/SC2/WG2
    document archives regarding the development of ISO 10646. But
    as Bob and Peter pointed out, this particular proposal has
    never reached a stage of formal ballotting in ISO.

    The Unicode Technical Committee also does post a little information
    about the simple cases where a proposal to encode one or more
    characters did not find agreement in the committee:

    http://www.unicode.org/alloc/rejected.html

    But the rationale for that page is mostly to keep people from
    independently coming back to again request encoding of something
    that was rejected (and would likely be rejected again), and for
    which a simple or obvious reason was available.

    There are many other instances, however, of proposals to the UTC
    which failed to attain agreement for one reason or another, which
    are *not* documented on that page, often because provision of a
    rationale to explain the situation would be too extensive to fit
    meaningfully into a simple HTML table cell entry. In extreme
    cases, there may not be *a* rationale at all -- simply many
    different objections of different grounds, leading to the inability
    to take any decision.

    The situation for diacritical extension of basic Arabic letters
    is a complex one involving decades of existing encoding practice
    at this point, long preceding the development of Unicode. It also
    involves complex issues of stabilization of of data under
    normalization, as well as model issues for the implementation of
    rendering. People on this list can discuss those issues further
    and explain why it is unlikely that the UTC would reconsider
    and take up a dynamic model for the encoding of Arabic letter
    diacritics, but that is probably all they can do at this point.

    If what you are looking for is a flawed, written rationale for
    a rejection which you can then appeal to a higher committee for
    reversal of decision -- well, you aren't going to find that,
    because that's not how this process works.

    Your task, instead, would be to create a consensus within the
    character encoding community (and the implementing information
    technology companies) that the existing Arabic encoding is so
    flawed that it requires introduction and implementation of
    a competing, distinct textual representation in Unicode.

    That, sir, is a *very* high mountain to climb, at this point.

    --Ken



    This archive was generated by hypermail 2.1.5 : Wed Jul 06 2005 - 19:13:04 CDT