Am 31.01.2008 um 16:08 schrieb Idris Samawi Hamid:

...

Any native arabic speaker out there? Is it %18 or 18% ??

It really depends on where you live. In Persian, 18% is common, among Arabs, %18. Both languages are left-to-right, but there is more than one convention for dealing with mathematical directionality.

Some of these issues are still being worked about in the Arabic-script world.

That's what I suspected ...

These subtlties need to be addressed for proper high-level handling of Arabic-script in ConTeXt/mkiv.

Yes!!

Best wishes Idris

Thanks, Steffen

Am 2008-01-31 um 16:08 schrieb Idris Samawi Hamid:

...

I have often seen (in arabic newpapers) the percentage sign set left to the number "% 18". Recently I read an article (http://sehstoerung.sonance.net/pdfs/ TitusNemeth_dissertation.pdf) that calls this "... the incorrect placement of the percentage sign".

Any native arabic speaker out there? Is it %18 or 18% ??

It really depends on where you live. In Persian, 18% is common, among Arabs, %18. Both languages are left-to-right, but there is more than one convention for dealing with mathematical directionality.

Some of these issues are still being worked about in the Arabic-script world.

These subtlties need to be addressed for proper high-level handling of Arabic-script in ConTeXt/mkiv.

Hm, reminds me of conventions where you place currency symbols - depending on if you live in a economist/accountant surrounding... Do you think some automated handling of such is possible at all? I know nothing about Arabic script, of course... Greetlings from Lake Constance! Hraban --- http://www.fiee.net/texnique/ http://wiki.contextgarden.net https://www.cacert.org (I'm an assurer)

Idris Samawi Hamid wrote:

On Thu, 31 Jan 2008 13:25:45 -0700, Henning Hraban Ramm wrote:

...

...

...

Any native arabic speaker out there? Is it %18 or 18% ??

...

...

These subtlties need to be addressed for proper high-level handling of Arabic-script in ConTeXt/mkiv. Hm, reminds me of conventions where you place currency symbols - depending on if you live in a economist/accountant surrounding...

Do you think some automated handling of such is possible at all?

One can already handle this sort of thing using Omega Translation Processes, which luaTeX supports. For example, there is an otp which makes sure that numeral-strings are always treated as LR, otherwise 1234 would come out as 4321. I suppose translating these rules into lua and/or opentype rules should be easy...

i can add this to the arab analyzer (just inject direction nodes around numbers) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

Peter Münster wrote:

On Thu, Jan 31 2008, Steffen Wolfrum wrote:

...

Any native arabic speaker out there? Is it %18 or 18% ??

Hello,

I don't know nothing about Arabic script, but wouldn't it be logic to ask for "%81 or 18%" ?

interesting point ... ok, small numbers, but how about [bara erom] 12.472.477 [bara emos] [bara erom] 774.274.21 [bara emos] for long numbers a pretty fast reversing mind is needed math goes from left to right but i once heard a talk about a tendency to go right - left as well (including mirrored symbols like \sum) ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

[bara erom] 12.472.477 [bara emos]

[bara erom] 774.274.21 [bara emos]

for long numbers a pretty fast reversing mind is needed

I think you're mistaken about the flexibility of human mind. Eight digits may look like a lot, but that's really not a long string compared to an average line width, and it does not take a lot of effort for the reader to look ahead for the start of the number. And eight digits is probably the longest it gets anyway (OK, maybe 12, but a 15-digit number would be difficult to read for anyone in any language). Actually you should look at it as a cultural difference, even if it interferes with scientific notation (and if you think about it, the percent sign is a scientific one, even if a rather simple and widely used one). And I'm sure you know how natural each person can find his own culture, while others would be puzzled by aspects of it. Hey, you're Dutch; doesn't that number read something like "twelve millions four hundred two and seventy thousands four hundred seven and seventy" in Dutch? (What else do German, Dutch, Arabic and Slovenian have in common?) And yet I'm sure you would read it out loud without hesitation (OK, you might say that you're only reversing two digits at a time, but I could reply that the way I see it, they are interspersed more or less arbitrarily).

math goes from left to right but i once heard a talk about a tendency to go right - left as well (including mirrored symbols like \sum)

That's standard in Maghreb as far as I know. Arthur

Hi all, I checked in some Persian printed books, and noticed that the rule is to write 18% or rather ۱۸% However I don't know of any book in Persian on rules of typography. But I am asking some people in Iran about the issue. Regarding the treatment of numbers, indeed numbers are written (and read…) in the usual way, but in XeConTeXt or in XeTeX (I don't know still how to typeset an Arabic or Persian file with mkiv LuaTeX), there is an issue with separators of digits: for instance if one writes (in the source file) ۱۲۳ ۴۵۶ ۷۸۹ (meaning 123 456 789, using a space as a separator between thousands) then one gets in the typeset file ۷۸۹ ۴۵۶ ۱۲۳ that is 789 456 123. To overcome this issue one may write \beginL ۱۲۳ ۴۵۶ ۷۸۹ \endL and then the output is correct, but there is a slight modification of the alignment and the treatment of glues (if I don't misinterpret). While if one uses a comma (that is U+002C, this is the comma for LR, while the comma for RL, the Arabic Comma, is U+060C) as in ۱۲۳{,}۴۵۶{,}۷۸۹ or ۱۲۳,۴۵۶,۷۸۹ or a dot (that is U+002E, which is indeed the same when writing LR or RL) as in ۱۲۳.۴۵۶.۷۸۹ then the output is correct. It should be feasible to have a mechanism allowing the user to enforce a certain type of separator between groups of digits, and this separator can be a space. Best regards: OK On 31 janv. 08, at 23:42, Hans Hagen wrote:

Peter Münster wrote:

...

On Thu, Jan 31 2008, Steffen Wolfrum wrote:

...

Any native arabic speaker out there? Is it %18 or 18% ??

Hello,

I don't know nothing about Arabic script, but wouldn't it be logic to ask for "%81 or 18%" ?

interesting point ... ok, small numbers, but how about

[bara erom] 12.472.477 [bara emos]

[bara erom] 774.274.21 [bara emos]

for long numbers a pretty fast reversing mind is needed

math goes from left to right but i once heard a talk about a tendency to go right - left as well (including mirrored symbols like \sum)

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl ----------------------------------------------------------------- ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : https://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

On 1 févr. 08, at 12:15, Steffen Wolfrum wrote:

[…]

Maybe I missed the point: don't both lines produce the same output: ۱۲۳ ۴۵۶ ۷۸۹ ?

\TeXXeTstate=1 \definedfont["GeezaPro" at 16pt]

\starttext ۱۲۳ ۴۵۶ ۷۸۹

\beginL ۱۲۳ ۴۵۶ ۷۸۹ \endL

\stoptext

... or was this a confusion of multi-script e-mails?

Hi Steffen, Indeed you are right. But the point is that when writing a text in Persian, or rather any RL text, in principle the beginning of the file is: \TeXXeTstate=1 \everypar={\setbox0=\lastbox \beginR \box0 } and therefore the output is different than the one you get. If one omits the second or both of the above two lines, again the output is as one intends, but the difficulty comes from the fact that (at least in Persian) when one writes some text containing a number, the digits and the separators, which can be the comma U+002C, the space, or the "slash" U+002F called Solidus (?) for writing fractions, should be typeset Left-to-Right. In particular compare the following for writing the number 123 456 789.01: ۱۲۳ ۴۵۶ ۷۸۹,۰۱ \beginL ۱۲۳ ۴۵۶ ۷۸۹,۰۱ \endL Indeed the first one results in something meaningless. Best regards: OK

On Fri, Feb 01, 2008 at 03:32:05PM +0100, Arthur Reutenauer wrote:

...

that is 789 456 123. To overcome this issue one may write

\beginL ۱۲۳ ۴۵۶ ۷۸۹ \endL

I think this is more or less what the Unicode bidi algorithm recommends (using the corresponding Unicode characters, of course, not e-TeX commands). I have to check.

Actually no, it expects you to use the non-breaking space (U+00A0 NO-BREAK SPACE) as a digit separator instead of the normal space. Sequences of digits interspersed with non-breaking spaces are not reordered by the Bidi Algorithm, contrary to what happens when you use a normal space. Of course this is not practical in ConTeXt for the moment, as the non-breaking space isn't treated as such (or is it?). Arthur

Idris Samawi Hamid wrote:

On Thu, 31 Jan 2008 15:35:35 -0700, Peter Münster wrote:

...

On Thu, Jan 31 2008, Steffen Wolfrum wrote:

...

Any native arabic speaker out there? Is it %18 or 18% ?? Hello,

I don't know nothing about Arabic script, but wouldn't it be logic to ask for "%81 or 18%" ?

Hmm, this could be a long discussion...

sure, but interesting (summary should be part of your arab typesetting paper)

In ancient times, Arabic numerals (whence our own) were pronounced as follows (translation):

1234 => "4 and 30 and 2 hundred and a thousand"

aha, makes sense maybe it's also related to the fact that nowadays we see more numbers, and have learned to see them 'as a whole' and not as a systematic sequence ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------