1. 19
  1.  

  2. 18

    Not this again.

    Different kinds of Arabic read numbers out loud differently, so some read most significant first, and some read least significant first, many are middle-endian. It’s pretty hard to make this conclusion just based on that.

    BUT: Arabic got the numerals from India. Indian languages write down numbers the same way you do in English, most significant digit first in a left to right script. Indian languages typically say numbers as most-to-least significant, with the middle-endian nature of tens and ones digits being swapped (similar to how “thirteen” is “backwards”). So it doesn’t even matter how it’s read in Arabic, because for a long time they were writing numbers the same way as the Indians, and it would be very weird if they used the same numerals but in the opposite order when looked at visually. The numbers got copied twice, from an LTR script to an RTL script back to an LTR script, without changing the order. The RTL nature of Arabic is a red herring here.

    And this doesn’t have much to do with why our computers work a certain way, anyway (see other comments).

    1. 2

      Different kinds of Arabic read numbers out loud differently, so some read most significant first, and some read least significant first, many are middle-endian. It’s pretty hard to make this conclusion just based on that.

      What do you mean by different kinds of Arabic? Classical and Modern Arabic both read numbers out loud the same, are you referring to dialectal Arabic? because most Arabs would not consider their dialect to be a legitimate Arabic, since almost all published text is written in Modern Arabic, or still in some cases Classical Arabic.

      For the record though this post is just wrong, because 521 in both Classical and Modern arabic is read as five hundred, one and twenty, and the author claims that “formal” arabic reads it differently. I’m not really aware of any dialects that read that as one, twenty, and five hundred, and at least to me (north African native speaker, but also Arabic language enthusiast) it sounds insanely wrong even for a dialect.

      BUT: Arabic got the numerals from India.

      Eastern Arabic Numerals are from India, Western Arabic numerals are believed to have originated in North Africa. European languages use the Western Arabic numerals, and not the eastern Arabic Numerals. The decimal system though, the more interesting part in my opinion, was indeed copied from Indian mathematicians.

      1. 2

        Eastern Arabic Numerals are from India, Western Arabic numerals are believed to have originated in North Africa

        Any sources for that? Most of the sources I have seen seems to trace the origins of both to Brahmi numerals.

        1. 1

          Right, dialectal Arabic, but also I knew that MSA doesn’t follow the same rules as what the author talks about and I wasn’t sure what they meant by “formal”, so I was really talking about MSA (I don’t know enough about Classical).

          Eastern Arabic Numerals are from India, Western Arabic numerals are believed to have originated in North Africa.

          No, the exact numeral forms were invented in North Africa (with influence from the eastern arabic forms!), but the concept of using numbers like that came from Eastern Arabic numerals, and Western Arabic numerals are a bit of an early divergence from them. They’re still related, see the first paragraph in https://en.wikipedia.org/wiki/Arabic_numerals#Origin_of_the_Arabic_numeral_symbols , in particular mathematicians had already trained themselves on Eastern Arabic numerals but hadn’t necessarily agreed on the forms.

      2. 11

        Well - interestingly, it has nothing to do with binary, bytes, or even computers - the problem comes from the fact that while our letters are Latin, our numerals are Arabic.

        Were that true, all of computers would be little-endian. The actual reason why Intel architectures became like that is in fact quite technical. Here’s a good answer on SO: https://stackoverflow.com/a/36263273 (tl;dr, to manipulate numbers wider than the CPU register without fitting them there completely, you have to start from the least significant machine word, which was 8 bit on 8008). Also note, that despite their “Arabic” nature, the bits within individual bytes still go from the most significant to the least.

        1. 9

          Also note, that despite their “Arabic” nature, the bits within individual bytes still go from the most significant to the least.

          I don’t think that’s observable, is it? The bits don’t have their own addresses, so we can’t say whether the high order bits have higher or lower addresses.

          1. 3

            Yes, that’s true! Thanks for the correction.

          1. 4

            This essay should also mention Germanendian, where the number “1521” is read as “five ten, one twenty”. In other words PDP-endian.

            1. 4

              521 […] a European reads it as “five-hundred, and twenty one”

              An English-speaking European, maybe. In German we spell it “fünfhunderteinundzwanzig”, which is literally “five-hundred and one and twenty”. Middle-endian, if you so wish.

              1. 5

                Oh man, don’t get me started on Danish.

                1. 1

                  French is almost as bad. I actually find Danish’s bonkers system easier to remember though for some reason.

                  1. 1

                    As an American, French is more annoying to vocalize (certain) numbers in than German. German was more: ok thats weird but its consistent and I don’t have to do math to figure out how to say the word. French on the other hand, made me very confused as to the history behind this. It seems to be the only romance language that does it too so I’m doubly confused.

                    1. 1

                      I agree, once you grok the vegisimal pattern it’s actually logical.

                  2. 1

                    which ironically is the exact way how it would be read in Modern Arabic and Classical Arabic. I’m really confused by the Quora answer that the OP based their thesis on. For Arabic speakers/readers here’s a great article that goes into the grammatical details, by a language scholar.