Re: On the possibility of encoding a portable interpretable object code into Unicode

From: John H. Jenkins (
Date: Thu Mar 26 2009 - 14:19:02 CST

  • Next message: Rick McGowan: "Re: On the possibility of encoding a portable interpretable object code into Unicode"

    First of all, I don't see why having a virtual machine programmed in
    as dedicated character code points is any better than having a virtual
    machine programmed via ASCII. It might be more compact, but from an
    implementation standpoint, that's about it's only advantage.

    To show what you're up against, just look over the history of Java and
    remember that it started out with a major player in IT pushing it.
    And Java distinguishes the text used to write programs from the
    portable byte code used to implement them. You, on the other hand,
    are going against fifty years' experience in CS of making a
    distinction between the text used to write programs and the machine
    language used to execute it.

    You'll have to convince people to write editors or compilers/
    assemblers, debuggers, and so on. You'll have to get a set of experts
    together to hammer out the details of the architecture and syntax.
    You'll have to determine how it will be displayed on systems that
    support it. What is supposed to happen if the user has a regular font
    installed covering these code points? How will the user distinguish
    cases where they want to *see* the code and cases where they want to
    *execute* it? What happens if they don't want to do either? What
    triggers execution, anyway? What happens when execution stops? What
    happens if execution *fails* to stop? What happens if the code
    attempts something illegal like dividing by zero? What will happen if
    this is embedded, say, in a URI or email or a file name? What happens
    if it's embedded in a word processing document? artwork? spread
    sheet? record in a database? instant message? Web page? comment on
    a blog? somebody's character name in an MMORPG? What are systems
    that don't support it supposed to do with it? How will you prevent
    people from using it to write viruses, trojan horses, and other
    malware? Where does input come from? Can it interact with a GUI?
    How do you display output? What is supposed to happen if you print
    it? Where is data stored? How will this interact with the various
    operating systems in existence? How does it interact with the
    applications on those systems? There are a hundred other questions
    that will need specific answers.

    *Plus* you'll have to convince people to use it, and enough of them
    that you reach critical mass.

    Once that is done, you'll have to show that there is a practical
    reason why this has to be implemented in Unicode and convince the UTC
    and WG2 that such encoding as plain text is the best solution to the
    practical problem(s) involved.

    By and large, any proposal for something which is not at least
    arguably "plain text" will not be favorably looked upon as a candidate
    for inclusion in Unicode. (N.B., I said "arguably". Personally, I
    don't think this is even arguably plain text and is, at best, a
    solution in search of a problem, but I don't want to resurrect any of
    the past wars we've seen on certain candidates proposed for
    encoding.) Unicode has, after all, more than once woken up with a bad
    hangover, looked at the character block lying next to it, and said to
    itself, "Did I *really* encode *that*‽‽"

    This archive was generated by hypermail 2.1.5 : Thu Mar 26 2009 - 14:21:32 CST