RE: BOCU-1 spec

From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Wed Feb 21 2007 - 16:58:00 CST

  • Next message: John H. Jenkins: "Re: Subj: MySQl and Unicode support"

    De : Doug Ewell [mailto:dewell@adelphia.net]
    > Has anyone else on this list actually read the BOCU patent? Among the
    > innovations claimed are:
    >
    > a. the method of encoding differences within the range -10FFFF hex to
    > +10FFFF hex.
    > b. the method of encoding U+0020 as itself to avoid frequent jumps.
    > c. the method "wherein the characters requiring higher code point
    > numbers are Greek" (i.e. obfuscated).
    >
    > This basically covers any method of encoding Unicode code points using
    > (a) a difference-encoding method, (b) a strategy to special-case certain
    > characters to improve efficiency, or (c) a system that encodes ASCII
    > characters using non-ASCII values. This is very far-reaching and could
    > be used against all kinds of alternative encoding schemes.

    I think it's wrong, because you read all claims as if one of them was just
    enough. The patent convers the case where the 3 claims are simultaneously
    used. If you remove one of these claims and want to pretend that this is
    BOCU, then what you define is not BOCU; any encoding using just claims a and
    c, or even just claim c would be blocked by the patent.

    If you just take claim c isolately, then it covers UTF-8, which is obviously
    prior art, but on fact ALL non-ASCII and non-ISO8859-1 encodings...

    If you just take claim b, then it covers several prior art where the space
    character is special as it does not change the current decoding state, and
    notably many text parsers.

    If you just take claim a, then it covers all UTF encodings (remember that
    UTF-32, for example can be thought as the encoding of differences, because
    there's in fat no formal difference between what is an "encoding" and what
    is a "decoding", notably when both are reversible to each other (i.e.
    objective). Claim a in fact is already completely covered by prior art (look
    for DPCM, i.e. differential pulse-coded modulation), whose patents have
    expired since long!

    I personally read this pattern as covering only products implementing these
    3 claims.



    This archive was generated by hypermail 2.1.5 : Wed Feb 21 2007 - 17:00:04 CST