Re: ISO 10646 compliance and EU law

From: Kenneth Whistler (kenw@sybase.com)
Date: Tue Jan 04 2005 - 17:40:20 CST

  • Next message: Mike Ayers: "RE: ISO 10646 compliance and EU law"

    Peter Kirk wrote:

    > I hate to throw a spanner in the works here, but from what I have been
    > led to believe the implication of the above is that a claim of
    > conformance to ISO/IEC 10646 is almost meaningless. For the Unicode
    > conformance clauses do not actually specify that the Unicode character
    > codes defined for a purpose must be used for that purpose. Or at least
    > so I have been assured. Let me explain.

    Or rather... obfuscate.

    >
    > In discussions a few months ago about a controversial proposed new
    > script,

    A cute way to reintroduce the argument about Phoenician to the
    list.

    > I was assured that acceptance of this new script did not imply
    > any obligation on existing users to use the characters of the new
    > script, rather than the characters of an existing closely related script
    > which were already being used for the new script by many users, with a
    > font change for the distinct character shapes. The situation would be
    > analogous to using Latin characters with a font change for writing
    > Greek, rather than the Unicode Greek characters.

    The analogy is bogus, as has been most of the discussion on this
    topic.

    If and when Egyptian hieroglyphics are encoded in the standard,
    nothing in the standard, whether in the conformance clause or
    elsewhere, will force anyone to use those characters for
    representing Egyptian texts. If they are used to using
    Latin *transliterations* for hieroglyphics, they can continue
    to do so, and nobody will stop them. Nor will their implementations
    suddenly become "nonconformant" or non-interoperable.

    If and when Phoenician characters are encoded in the standard,
    nothing in the standard, whether in the conformance clause or
    elsewhere, will force anyone to use those characters for
    representing Phoenician (or Punic or Ancient Hebrew) texts. If they
    are used to using Modern Square Hebrew *transliterations* for
    Phoenician (or Paleo-Hebrew), they can continue to do so, and
    nobody will stop them. Nor will their implementations suddenly
    become "nonconformant" or non-interoperable.

    > I was assured that
    > there was nothing in the Unicode conformance clauses which prevented
    > users from continuing to use such masquerading font solutions, which
    > have of course been in existence since long before Unicode.

    This "masquerading font solution" is a transliteration masquerading
    as a masquerade.

    Whether a particular transliteration can be *implemented* in
    terms of a simple font shift depends on the nature of the
    transliteration. In the case of Canaanite and Modern
    Square Hebrew, well, you can set up a nice one-to-one mapping
    of 22 letters and hack away with your fonts. In the case
    of Egyptian hieroglyphics and Latin, the transliteration
    relationship is too complicated to do a simple mapping for,
    so nobody even attempts to do this with just a font mapping.
    Instead, a full transliteration scheme is put in place,
    together with a markup syntax, to enable shifting from one
    to the other for a rendering.

    > So, would a system which consistently used the codes for Latin
    > characters for writing the Greek alphabet, and distinguished Greek
    > script from Latin by a font change, conform to Unicode, and to ISO/IEC
    > 10646?

    It depends, of course.

    If the implementer was confused about the distinction between
    Latin and Greek, and thought that they were the same characters
    with a font shift, then no.

    If the implementer deliberately created a convention for
    the representation of Greek using the Latin alphabet, and
    provided a font mapped accordingly, published his conventions,
    and encouraged you to buy his new Graecolatin Deluxe Tool,
    then yes. Although he would have a hard time getting people
    to buy into his peculiar text conventions, however.

    Nothing in the standard prevents people from using it
    to create arbitrary ciphers, nor even wackos from coming
    up with off-the-wall transliterations or worse from one
    form of text to some other form. Or from implementing
    the same by means of silly mappings in fonts. If you share
    their wacko text conventions, then you can use them
    to your heart's content, while still using the standard
    conformantly. If I want to send you Biblical texts
    represented in a Yi cipher that gets rotated on a daily
    basis, nobody is going to stop me or you from engaging
    in such silliness.

    > And would it be acceptable to the EU?

    Rhetorical question, right?

    > I suspect that the EU wants
    > to specify something stronger, that the actual ISO/IEC 10646 Greek
    > characters are used for writing Greek script. But maybe, if they have in
    > fact specified something, it is much less than they intended.

    Really, Peter. What is the point of taking something that
    is actually quite simple and trying to convince everyone
    that it is arcane and unanswerable?

    --Ken



    This archive was generated by hypermail 2.1.5 : Tue Jan 04 2005 - 17:43:57 CST