In Algeria, Kabyle is taught in the latin alphabet (schools/university). All experts in the kabyle language use the latin alphabet (mouloud mammeri, salim chaker, inalco, académie berbère...etc), and all kabyle websites who write in kabyle write it in the latin alphabet. (www.imyura.net or tamazight.fr for example)

All those who are willing to participate have studied the grammar of the kabyle language in the latin alphabet (including me), so obviously the kabyle wiki will be written in latin.

But if you want me to develop some arguments on why we prefer the latin alphabet over the tifinagh or cyrillic or chinese or hebrew or eventually arabic alphabet, here they are:

-We are a little people of 3 millions in Kabylie (6 if we add the diaspora), so we choosing the most practical solution : the latin alphabet. Just have a look at the kurdi wikipedia which uses the latin and arabic alphabet, 99% of the articles are in the latin alphabet, and less than 1% are in the arabic script. Also note that the number of kurdi-speaking people is the tenth of that of the arabic-speaking people but the kurdi wiki have more articles than the arabic wiki, add to this paradox the fact that 99% of kurds never studied their language because the states where kurds live do not promote the kurdi-language. facts prove that the latin alphabet is much more practical and efficient.

-Modern Tifinagh works exactly like the latin alphabet, it is written from left to right and has vowels, so it is easily convertible to latin with the use of a software (it would take less than one second for the whole process). but as I said earlier we seek efficiency in both real world and internet. No one's going to read the articles in the tifinagh alphabet... and especially write them on a paper. I have an experience of 5 years writing with the tifinagh alphabet, but still, it takes about 5 seconds to write a word in tifinagh but less than 1 second in latin. (don't forget that tifinagh dates back to 500 BC, so the letters are very primitive, very big, and complex to write.. I'd say draw since they look like geometric figures). Trying to improve the tifinagh would mean creating a new alphabet, sincerely we do not have the capacity. Also, "improving" means that it (tifinagh) is used, which is not the case.

-As for the rest of the remaining scripts, I wonder why we would use them since we don't even use the one our ancestors used. Moreover, 99% of newly written languages are in latin alphabet, that's because it is the most efficent. Do not get me wrong, I respect the chinese who wants to keep using the chinese pictograms, it's good that they keep an ancestral tradition but I'm not going to adventure my language which is very, very weak into more troubles.

-some say that because Berber works like other afro-asiatic language, by knowing the grammar and some practice, you can guess 90% of the words' vowels, therefore it should use the alphabet used by other afro-asiatic languages. This is also the case for english :

"Many non-Semitic languages such as English can be written without vowels and read with little difficulty. For example, the previous sentence could be written"

"Mny nn-Smtc lnggs sch `s `nglsh cn b wrttn wtht vwls `nd rd wth lttl dffclty." (example taken from wiki)

but I care a lot for the remaining 10%. the second sentence looks pretty much primitive if you ask me. Also, not all afrasian languages use hebrew or arabic alphabet, for example maltese and ethiopian use the latin alphabet. Speaking about maltese... there 1000 times more arabs than maltese (300 millions against 0.3 million), but the maltese wiki has half the number of what the arabic wiki has as articles.

-I hope I have answered your main question, and most importantly anticipated your eventual replys.

Neil Harris <usenet@...> writes:

Given your arguments above, it looks like a no-brainer to start the Kabyle wiki in the Latin script.

However, if there is, as you say, a deterministic transformation to/from Tifinagh, we could easily later use a variant of the multi-script software currently used to handle the different Chinese scripts on the Chinese Wikipedia to implement a dual-script interface, or at the least an alternative Tifinagh display mode for the Latin-based text.

Do you have any references for Tifinagh <-> Latin transliteration?

You mean no-brainer to start it in Tifinagh? (you've written latin), Yeah I guess so.

The link to the transliteration of tifinagh could be found in wikipedia:

http://en.wikipedia.org/wiki/Tifinagh

But there is a probleme with the transliteration in latin, it's not the one used to write Kabyle (we actually use a modified latin alphabet).

I'll try to add do it here :

http://en.wikipedia.org/wiki/Berber_alphabet

On 10/27/06, Mark Williamson node.ue@gmail.com wrote:

"No-brainer" is a sort of idiom. Paradoxically, it means "the obvious choice".

I.e., one you don't need brains to make correctly, not one that you would only make if you had no brains.

Neil Harris wrote:

Thanks for the link in the [[Berber alphabet]] article: given that reference, this Latinized version of the Berber alphabet looks like it should be able to be round-tripped to/from the Tifinagh script.

I'm only confused by one thing: the article says that the two non-Latin characters are Cyrillic, but both of the characters listed in the tables themselves appear at first sight to be Greek letters (epsilon and gamma respectively).

(See also http://www.win.tue.nl/~aeb/natlang/berber/tifinagh/tifinagh-ircam.html)

However, I think they might be intended to be

U+025B LATIN SMALL LETTER OPEN E (which looks like an epsilon)

and

U+0263 LATIN SMALL LETTER GAMMA (which looks like a gamma)

respectively. This would also have the effect of making the Latinized Berber alphabet much easier to work with, since every character in it would then be within either the Unicode Latin or IPA extensions ranges, and thus fall entirely within the Latin writing system.

-- Neil

If the above is the case, then the Unicode translation table, based on the [[Berber alphabet]] article, would then look like this:

Latin Tifinagh

U+0041 LATIN CAPITAL LETTER A <-> U+2D30

U+0042 LATIN CAPITAL LETTER B <-> U+2D31

U+0043 LATIN CAPITAL LETTER C <-> U+2D5B

U+0044 LATIN CAPITAL LETTER D <-> U+2D37

U+0045 LATIN CAPITAL LETTER E <-> U+2D3B

U+0046 LATIN CAPITAL LETTER F <-> U+2D3C

U+0047 LATIN CAPITAL LETTER G <-> U+2D33

U+0048 LATIN CAPITAL LETTER H <-> U+2D40

U+0049 LATIN CAPITAL LETTER I <-> U+2D49

U+004A LATIN CAPITAL LETTER J <-> U+2D4A

U+004B LATIN CAPITAL LETTER K <-> U+2D3D

U+004C LATIN CAPITAL LETTER L <-> U+2D4D

U+004D LATIN CAPITAL LETTER M <-> U+2D4E

U+004E LATIN CAPITAL LETTER N <-> U+2D4F

U+0051 LATIN CAPITAL LETTER Q <-> U+2D47

U+0052 LATIN CAPITAL LETTER R <-> U+2D54

U+0053 LATIN CAPITAL LETTER S <-> U+2D59

U+0054 LATIN CAPITAL LETTER T <-> U+2D5C

U+0055 LATIN CAPITAL LETTER U <-> U+2D53

U+0057 LATIN CAPITAL LETTER W <-> U+2D61

U+0058 LATIN CAPITAL LETTER X <-> U+2D45

U+0059 LATIN CAPITAL LETTER Y <-> U+2D62

U+005A LATIN CAPITAL LETTER Z <-> U+2D63

U+010D LATIN SMALL LETTER C WITH CARON <-> U+2D5B

U+1E0D LATIN SMALL LETTER D WITH DOT BELOW <-> U+2D39

U+011F LATIN SMALL LETTER G WITH BREVE <-> U+2D35

U+1E25 LATIN SMALL LETTER H WITH DOT BELOW <-> U+2D43

U+1E63 LATIN SMALL LETTER S WITH DOT BELOW <-> U+2D5A

U+1E6D LATIN SMALL LETTER T WITH DOT BELOW <-> U+2D5F

U+1E93 LATIN SMALL LETTER Z WITH DOT BELOW <-> U+2D65

U+025B LATIN SMALL LETTER OPEN E <-> U+2D44

U+0263 LATIN SMALL LETTER GAMMA <-> U+2D56

Presumably, since there is no case distinction in Tifinagh, both the upper and lowercase versions of the Latin letters should map onto the same Tifinagh letters. Is there a particular preference about the mapping in the opposite direction?

Certainly most Latin-alphabet readers would find all-lower-case preferable to all-upper-case for comfortable reading, and the table at http://www.win.tue.nl/~aeb/natlang/berber/tifinagh/tifinagh-ircam.html also suggests that lowercase was intended to be the standard.

-- Neil

Except that I forgot to hand-patch the correct Latin characters in at the end, last time: corrected, this should have read:

U+0061 LATIN SMALL LETTER A <-> U+2D30 TIFINAGH LETTER YA

U+0062 LATIN SMALL LETTER B <-> U+2D31 TIFINAGH LETTER YAB

U+0063 LATIN SMALL LETTER C <-> U+2D5B TIFINAGH LETTER YASH

U+0064 LATIN SMALL LETTER D <-> U+2D37 TIFINAGH LETTER YAD

U+0065 LATIN SMALL LETTER E <-> U+2D3B TIFINAGH LETTER YEY

U+0066 LATIN SMALL LETTER F <-> U+2D3C TIFINAGH LETTER YAF

U+0067 LATIN SMALL LETTER G <-> U+2D33 TIFINAGH LETTER YAG

U+0068 LATIN SMALL LETTER H <-> U+2D40 TIFINAGH LETTER YAH

U+0069 LATIN SMALL LETTER I <-> U+2D49 TIFINAGH LETTER YI

U+006A LATIN SMALL LETTER J <-> U+2D4A TIFINAGH LETTER YAZH

U+006B LATIN SMALL LETTER K <-> U+2D3D TIFINAGH LETTER YAK

U+006C LATIN SMALL LETTER L <-> U+2D4D TIFINAGH LETTER YAL

U+006D LATIN SMALL LETTER M <-> U+2D4E TIFINAGH LETTER YAM

U+006E LATIN SMALL LETTER N <-> U+2D4F TIFINAGH LETTER YAN

U+0071 LATIN SMALL LETTER Q <-> U+2D47 TIFINAGH LETTER YAQ

U+0072 LATIN SMALL LETTER R <-> U+2D54 TIFINAGH LETTER YAR

U+0073 LATIN SMALL LETTER S <-> U+2D59 TIFINAGH LETTER YAS

U+0074 LATIN SMALL LETTER T <-> U+2D5C TIFINAGH LETTER YAT

U+0075 LATIN SMALL LETTER U <-> U+2D53 TIFINAGH LETTER YU

U+0077 LATIN SMALL LETTER W <-> U+2D61 TIFINAGH LETTER YAW

U+0078 LATIN SMALL LETTER X <-> U+2D45 TIFINAGH LETTER YAKH

U+0079 LATIN SMALL LETTER Y <-> U+2D62 TIFINAGH LETTER YAY

U+007A LATIN SMALL LETTER Z <-> U+2D63 TIFINAGH LETTER YAZ

U+010D LATIN SMALL LETTER C WITH CARON <-> U+2D5B TIFINAGH LETTER YASH

U+1E0D LATIN SMALL LETTER D WITH DOT BELOW <-> U+2D39 TIFINAGH LETTER YADD

U+011F LATIN SMALL LETTER G WITH BREVE <-> U+2D35 TIFINAGH LETTER BERBER ACADEMY YAJ

U+1E25 LATIN SMALL LETTER H WITH DOT BELOW <-> U+2D43 TIFINAGH LETTER YAHH

U+1E63 LATIN SMALL LETTER S WITH DOT BELOW <-> U+2D5A TIFINAGH LETTER YASS

U+1E6D LATIN SMALL LETTER T WITH DOT BELOW <-> U+2D5F TIFINAGH LETTER YATT

U+1E93 LATIN SMALL LETTER Z WITH DOT BELOW <-> U+2D65 TIFINAGH LETTER YAZZ

U+025B LATIN SMALL LETTER OPEN E <-> U+2D44 TIFINAGH LETTER YAA

U+0263 LATIN SMALL LETTER GAMMA <-> U+2D56 TIFINAGH LETTER YAGH

-- Neil

Toira wrote:

Thanks for the correction mark ;-)

Thanks for spending so much time Neil :-)

You're welcome.

Yes I agree with your choices, the conversion from tifinagh to latin should be done all in lower case (otherwise it would be a pain in the eyes).

BTW, I've corrected the mistake about cyrillic, I apologise for my ignorance, I have replaced it with greek alphabet (epsilone and gamma are greek letters).

Just one question Neil, are you actually developping the kab wiki? because I did not get it yet.

No, I'm not. But you are posting to the right mailing list for dealing with technical questions about multi-script Wikipedia editions.

I'm sure someone here can point you at the right place for formally requesting the creation of a new Wikipedia. They might want to hear if you have any other people ready to edit it, or if you are willing to set up a small trial Wiki in that language before going live with a full kab Wikipedia.

I suggest that you ask for the kab Wikipedia to be set up in the Latin script at first, because that's quite the easiest option, but make clear that you:

a) later want to use the multi-script interface software, and b) that it is a simple transliteration issue, rather than a complex set-up like Chinese, and that c) you already have Unicode assignments and a transliteration table ready to go...

And oh please, can you give me the link to the chinese wiki? just to see how the conversion work.

Toira.

Chinese Wikipedia (main entry point):

http://zh.wikipedia.org/

Chinese Wikipedia main page in traditional Chinese script (as used in Taiwan):

http://zh.wikipedia.org/w/index.php?title=%E9%A6%96%E9%A1%B5&variant=zh-...

Chinese Wikipedia main page in simplified Chinese script (as used in mainland China):

http://zh.wikipedia.org/w/index.php?title=%E9%A6%96%E9%A1%B5&variant=zh-...

Regards,

Neil