# 4.0699 Bi-Directional Lines (2/71)

Elaine Brennan & Allen Renear (EDITORS@BROWNVM.BITNET)
Wed, 7 Nov 90 11:43:23 EST

Humanist Discussion Group, Vol. 4, No. 0699. Wednesday, 7 Nov 1990.

(1) Date: Sun, 4 Nov 90 16:05:08 -0600 (36 lines)
From: Alan D Corre <corre@csd4.csd.uwm.edu>
Subject: direction of numbers

(2) Date: Sun, 4 Nov 1990 14:50:11 GMT+0300 (35 lines)
From: Judy Koren <LBJUDY@VMSA.TECHNION.AC.IL>
Subject: RE: 4.0690 R: Arabic Word Processing

(1) --------------------------------------------------------------------
Date: Sun, 4 Nov 90 16:05:08 -0600
From: Alan D Corre <corre@csd4.csd.uwm.edu>
Subject: direction of numbers

Patrick Connor's comment "it [the Arabic WP] must run right to left for
text and then run left to right for numerals" reveals an interesting
piece of ethnocentrism in which most of us are enmeshed. I would assert
that our numbers do, in fact, run from right to left. When you write a
number from left to right, the first digit is ambiguous until the number
is finished (watch what happens when you enter a sum on your electronic
bank machine.)

If we read numbers the other way around, this ambiguity would not exist.
Think of it this way. When you write the numbers from 1 to 9, the logical
thing is to have this position revert to 0, and then write the new tens
digit after it, i.e. to the right. So ten would be written 01.

I assume that our habit of writing numbers "backwards" is because we took
over this habit from the Arabs, whose contributions were so important in
the development of mathematics.

An interesting sidelight. Alan Turing probably was unaware of all this,
but being the brilliant man he was, sensed that our numbers are the
wrong way round, for people who are in the habit of writing from left to
right, and insisted that binary numbers at least should go from left to
right, so in 10010011 the *leftmost* 1 was to be understood as unity.
An individual who attended a lecture he gave records that he did *all*
of his arithmetic on the blackboard the "other way around" to the total
befuddlement of his audience. I suspect that he was not showing off or
trying to confuse. It just didn't occur to him that not everyone shared
his talent for ambidirectional arithmetic.

Alan D. Corre
Department of Hebrew Studies
University of Wisconsin-Milwaukee (414) 229-4245
PO Box 413, Milwaukee, WI 53201 corre@csd4.csd.uwm.edu
(2) --------------------------------------------------------------41----
Date: Sun, 4 Nov 1990 14:50:11 GMT+0300
From: LBJUDY@VMSA.TECHNION.AC.IL
Subject: RE: 4.0690 R: Arabic Word Processing

It isn't that difficult to achieve right-to-left (for Arabic) and
left-to-right (for English or numerals) in the same line; all the
Hebrew word processors do it routinely on DOS machines. A very popular
word processor for Hebrew over here is Wordmill; a friend of mine uses
an Arabic version of it (i.e. Arabic/English rather than
Hebrew/English). As far as I know it chooses the correct forms of the
letters, can of course enter both languages on the same line, or Arabic
+ numerals, etc. Though the problem with a bi-directional line is to
divide the line correctly if you reach the end of the line in the
secondary language, i.e. the one that runs in the other direction from
most of the paragraph. The new line will start the other side from the
side native to the language you're typing (e.g. if in an Arabic
paragraph you insert a few words of English and the English phrase
overruns the line, the new line will place the last word(s) of the
English insert at its right-hand end). Somehow it never looks right.

Another problem is that Hebrew word processors in Israel, and presumably
Arabic ones in Arab countries, presuppose that the alphabet in question
is in RAM; obviously most U.S. users don't have anything but the IBM
character set in RAM. EGA-and-up users don't have this problem because
the character set is a soft font, not RAM-based. I seem to recall that
my friend with the Arabic version of Wordmill uses a Hercules Plus card
(now long off the market), but presumably Arabic Wordmill must have font
support for EGA and up too. I tried Wordmill (in Hebrew) three versions
ago and was not impressed. It's now, so they tell me, much more
user-friendly and also more flexible. At any rate it's got a good 80%
of the market over here, I estimate, but I suspect that's because it
concentrates (like WordPerfect?) on the business side rather than the