machine code language

From Lojban
(Redirected from Machine code language)
Jump to navigation Jump to search

mi'e tinkit just an idea i've had that if we make a language that is not restrained at all by letters and verbalization, but rather is specified numerically, much like computer machine code.

Go ahead and do it, then! (Just don't call it Lojban.)

ok i'll call it 1010110101 .u'i --tinkit

Quoted without permission from [1]:

Anyway, telegram operators were so skilled that they were known to

chat using 4 digit numerals with each other, in front of the nosy eyes of

guess-whos. You could eavesdrop onto a not very philosophical dialogue

between two buddies, Misters Gao and Sima.

S: (slapping G on the back) "3674,2998,6745,2007,6683,0123!"

G: (Apologetically..) "1322,1322,6453,0851,9052....."

S: (Chuckling) "4390,2391,3255,0038,1091."

Both: "4598,3411,4598,3411!" (Giggling wildly... no codes available).

-mi'e tsali

I'd have to look it up again, but I recall something similar to that from way back in the 1600s(!) by a man named Lodowick... Argh, it's not in my head anymore. I have to find the book I had, and it's packed in some box someplace. Basically a Dewey Decimal System for ideas that are words. The conlang Ro had a similar notion, though based on sounds and not explicitly numbers, and there was some of that in Babm too. I think EmSighAy actually did use numbers, and was all mathematically defined. --mi'e mark

Can language be specified purely numerically?

47 65 65 2c 20 49 20 64 6f 6e 27 74 20 6b 6e 6f 77 2e

ru'a that's ASCII code. totally different--the language is created for human pronunciation and translated to numbers. this would be created to be encoded in numbers, and pronounced like "panopanononopanonopanopapano" .u'i --mi'e tinkit

your string of "pa"s and "no"s contains exactly 12 bits of information if I could correctly. the phrase "wooly wild weasel" (half as many syllables) contains about 4 times as many bits of information. (google for information theory) what use is something with an information density 1/8th of normal languages? ru'a no te pilno --jay

  • the object wouldn't be high density of information--otherwise the rules could get very complicated. just the chance to see how language issues are treated when human pronunciation is not an issue at all.

(mi'e keidarien) Since Lojban has digits for hexadecimal, saying binary isn't really that sparse. But still, it's kind of like saying "four seven six five six five two cee ..." rather than "onezerozerozerooneonezerooneonezeroonezeroone...". Speaking binary is just silly, and not terribly useful. After all, one of Lojban's great strengths is that it uses (almost) all the sounds a human mouth can make in a very efficient manner.

It may be useful to use this kind of binary lojban for natural-language compression. The difficulty then is that lojban words are designed around the 26 character alphabet (a to z, except hqw, adding '.,) and there's no simple mapping from one to the other.

What is this good for?? Okay, for communication between machines! This is what computers are all based on. Yet, never between men, albeit there's a special goal e.g. encryption! Think of WWII Navajo Code Talkers: Here, generally, each Navajo word was to give only one single letter (i.e. the first one of the words equivalent in English). Principally, a bad density ratio (yet, there also were shortcuts).

  • Ok, great idea, but there's a difference here: Navajo Code Talkers worked because Navajo was a totally obsurce language, not a normal encryption like had been previously used. (Not that Lojban isn't totally obscure...) Btw, the Navajo Code Talkers didn't use a word-to-letter mapping, (does navajo even use a-z?) they had word-to-word, for example, a "turtle" meant "Tank", "Big Man" meant "the President", and lots others. I think they used regualar Navajo words for normal stuff, but again, the reason this worked it because the Japanese didn't have a clue about Navajo. --keidarien
    • Navajo has a-z (I forget what letters are actually in use) and some additional items for peculiarly Navajo (Dine') sounds. That is all by the way; the characterization of Code Talkers is correct.
      • As mentioned, they used a word-to-letter mapping - and lots of shortcuts (word-to-word mapping): have a look here [2] - and BTW, there was no need for having letters like in English because they actually used the English equivalent words initials ;-)