Re: Hexatridecimal

From: Philippe VERDY (
Date: Thu Jun 16 2005 - 13:44:17 CDT

  • Next message: Asmus Freytag: "RE: It is easy to predict the past."

    hexatridecimal notation for numbers does not look like a notation used for actual languages or culture. It certinaly has applications, but it looks too much like a technical convention, detached of any localization. So I really wonder if it is even important that this notation becomes localizable, given that it already uses ththis base simply because the ASCII alphabet is limited to 26 letters. Many European languages have more letters (including those with accents), but I have never seen any numeric system using it. The fact that these 26 base letters are universally supported in Latin-based languages is enough to limit the subset out of any languistic consideration.
    Numeric systems also don't like much the natural property of text to use ligatures, and variable width characters (there are known exceptions in Indic numeric systems). For these reasons, the choice of the 26 letters is very much arbitrary. I would not expect this convention to be "ported" easily to other scripts (for example Hebrew or Arabic, due to the difference of directionality between letters and digits...). A more significant example would be to port this numeric system to the Greek or Cyrillic alphabets (but then, what is the basic alphabet suitable for numeric systems? Don't they have a different number of base letters? What will happen if the basic set of letters has less than 26 letters in these scripts?)

    Hummm... Try to convince us that numeric systems mising digits and letters also exist and are readable and suitable for technical conventions. If these conventions exist, they need a stable alphabet, with an unambiguous and wellknown order. Not so many people even know the correct order of the first 3 letters in the Greek and Cyrillic alphabets.
    This question will be even more difficult with syllabaries (for example Katakana). What will be really convincing is the fact that there's effectively a cultural convention for numeric systems using a stable set of letters or digits to represent numbers. Unicode has encoded some of them which are not using the decimal positional system: for example the Roman numeric letters, but also old Tamil numbers (before the adoption of a digit for zero and use of the positional system).
    So I don't care much of the technical only base-36 system (10 figures + 26 uncased Latin letters) is localizable. Same thing for other technical conventions (Base32, Base64, etc...) These conventions are FIRST defined by their set of digits. These digits are not localizable, because their locale is intrinsicly the numeric convention itself. Changing these digits by such "pseudo-localization" would break those systems the same way as if you attempted to use Indo-Arabic digits in Latin texts: you would violate the cultural locale convention.
    So don't speak about set of digits. The localization of numbers is FIRST the selection of the appropriate numeric system used in a culture. The Base-36 system is not part of any modern or old humane locale. So no need to "translate" it.
    > Message du 16/06/05 19:44
    > De : "Mark E. Shoulson"
    > A : "JFC (Jefsey) Morfin"
    > Copie à :
    > Objet : Re: Hexatridecimal
    > Check into "Quoted-Printable" format, also base64 encoding, UTF-7, and
    > uuencode for that matter. People have been using formats of that sort to
    > encode non-ASCII text for quite some time.
    > ~mark
    > JFC (Jefsey) Morfin wrote:
    > > Would someone knows if work has been carried on the portability in
    > > different scripts of hexatridecimal, i.e. numbers figured by 0-Z in
    > > ASCII (10 figures + 26 characters). Treating ASCII sequences in
    > > protocols and programs as numbers could simplify protocol
    > > multilingualisation.
    > > thank you.
    > > jfc
    > >

    This archive was generated by hypermail 2.1.5 : Thu Jun 16 2005 - 13:51:18 CDT