- From: Youcef Rabah Rahal <y.rahal@gmail.com>
- Date: Wed, 10 Nov 2004 16:20:08 +0100
- To: www-math@w3.org

Hello, I would like to give my contribution on the subject. Some of the points below have already been discussed, but I'll discuss them again in order to emphasize their importance. In order to prepare the specification concerning maths in Arabic, it's important to note the following: 1- The user should be able to use either 'Arabic' or 'Hindic' numerals. Writing maths in Arabic does not mean necessarily using Hindic digits ('Arabic' numerals have been used hitstorically in the western part of the Arab world, called Maghreb, for centuries now). 2- The specification should consider that maths of _any_ level can be written in Arabic. Indeed (I can speak only of Algeria), we use Arabic notation from elementary school (apples and oranges :-) until the last year of high school. The notation is used in maths, physics, chemistry etc. Although, in university, LTR notation is used (but maybe not for long). Since high school maths are written in Arabic, I believe we can assume the RTL property. 3- If a transliteration map is to be considered, one must keep in mind that variable notations may vary. For example, in Algeria, the y variable is noted 'Ayn' (U+0639). In the Mashreq (eastern part of the Arab world), the letter 'Sad' is used U+0635... And vice-versa, for the z variable ! I don't know if other countries of the Maghreb have also this particularity. Pr. Lazrek ? 4- Same as the point above. To handle the functions, it's good to know that function abbrevations are not quite the same between Algeria and the Mashreq (again, I need a confirmation to say if the case of Algeria can be generalized to the whole Maghreb). So, even if the _names_ of the functions are the same, trigonometric functions like cosine, don't have the same symbol: noted "JATA" in Mashreq and "TAJIB" in Maghreb (or only in Algeria ?), though it has the same name "Jayb Tamam". 5- This has already been said, but it's important to say it again :-) Most of the matematical symbols need to be mirrored (well, except of the symmetrical ones, that is...). This includes, sqrt, sum, integral, inclusion, limits etc etc. 6- This also has already been said and I don't know if this is the proper channel to discuss it. But it's important to note that the Unicode statement which says that numbers in Arabic are written LTR has to be discussed at some point. Numbers in Arabic are read _and_ written from RTL. Even if in the few last years a new way of writing and reading numbers has emerged (mainly due to the use of LTR software ?). It's not too late to fix that however. Now a question: is there any deadline on the document preparation ? And who will be in charge of its preparation ? Regards, PS: A little about me: I'm from Algeria and have a recent (Jan 2004) PhD in Physics (Observational Cosmology) from Pierre & Marie University, Paris. I am also a member of the Arabeyes.org community (www.arabeyes.org) and work on a scientific/symbolic computation and graphic sofftware called Bayani (which is still in its early development stage) and which works indifferently in any RTL or LTR language, so I'm particularly interested in MathML i18n :-) -- Dr. Youcef R. Rahal

Received on Wednesday, 10 November 2004 15:28:11 UTC