[ih] Erratum in IEN 37
Carsten Bormann
cabo at tzi.org
Wed Apr 7 08:32:54 PDT 2021
OK, I already sent a version of this off-list, but if people really want to discuss this here...
What Brian called Arabic numerals actually are Arabic-Indic (or Eastern Arabic) numerals [1], which are handled exactly like Western Arabic numerals (except that they are harder to see in a hex-dump like below).
From Arabic Wikipedia:
بداية الحرب الباردة في 1962 والتي انتهت في 1979.
=
The beginning of the Cold War in 1962, which ended in 1979.
(Something moved around the period to the right in my mail reader; no idea what you will see.)
00000000: d8a8 d8af d8a7 d98a d8a9 20d8 a7d9 84d8 .......... .....
00000010: add8 b1d8 a820 d8a7 d984 d8a8 d8a7 d8b1 ..... ..........
00000020: d8af d8a9 20d9 81d9 8a20 3139 3632 20d9 .... .... 1962 .
00000030: 88d8 a7d9 84d8 aad9 8a20 d8a7 d986 d8aa ......... ......
00000040: d987 d8aa 20d9 81d9 8a20 3139 3739 2e .... .... 1979.
Or a different sentence in Persian, with Persian-Arabic numerals (where 4, 5, 6 look different), with annotation added:
"سال ۱۹۶۲ (هزار و نهصد و شصت و دو) میلادی، دومین سال از دههٔ ۱۹۶۰ در سدهٔ ۲۰ میلادی بود
>> ".codepoints.map { "%x" % _1 }.join(" ")
=> "633 627 644 20 **6f1 6f9 6f6 6f2** 20 28 647 632 627 631 20 648 20 646 647 635 62f 20 648 20 634 635 62a 20 648 20 62f 648 29 20 645 6cc 644 627 62f 6cc 60c 20 62f 648 645 6cc 646 20 633 627 644 20 627 632 20 62f 647 647 654 20 **6f1 6f9 6f6 6f0** 20 62f 631 20 633 62f 647 654 20 6f2 6f0 20 645 6cc 644 627 62f 6cc 20 628 648 62f a"
>>
So numbers actually *are* encoded big-endian in arabic script; we see them in our familiar writing direction (i.e., inverse to the text) because they are LTR fragments in RTL script.
Grüße, Carsten
[1]: See also https://en.wikipedia.org/wiki/Arabic_numerals
More information about the Internet-history
mailing list