@Amastan

thanks for your message.

We launched series of localizations and NLP tools for the kabyle language only and we can't manage all flexions of the whole language berbers with one corpora. You know, every berber dialect has it's own functions, derivational morphems and inflectional morphemes, grammar system. The system of these morphemes (derivational/inflectional) make changes to the meaning of the stem and the grammatical information for each berber language. Pronounciation as well.

Yes I know that a lot of words are shared between them beacause they belong to the same family like latin languages, but there is a gap within pronounciation, grammar structure..... and so on.

Iso 639-2 and Iso 639-3 listed all the codes for each language. Some of them are listed on iso 639-2 such as: Kabyle (kab), Tamasheq (tmh), Standard Moroccan Tamazight (zgh), please see http://www.loc.gov/standards/is.../code_list.php and the others listed on iso 639-3 such as Mozabit, Chawi, ...and the others : please see : https://iso639-3.sil.org/code_tables/639/data Example: for the mozabit the code page can be reached here: https://iso639-3.sil.org/code/mzb.

We also launched Kabyle language on CLDR (Common local Data Repository, a project of the UNICDE Consortium, please see: http://st.unicode.org/cldr-apps/v#locales/kab), the project is now maintained under the "kab" code (Kabyle language). Another team is maintaning the kabyle language on Wikimedia (wikipedia and wikitionary: please see https://kab.wikipedia.org/ and https://kab.wiktionary.org , wikitionary is under incubation). 27 Mozilla projects are also localized into the kabyle language: Firefox for all plateforms, Tunderbird,... are now available in kabyle language, ... I can't list all the projetcs we launched for the kabyle language (LibreOffice, OpenOffice, LinuxMint.... and some others are comming in kabyle language).

Anyway, we are know launching series of NLP tools dealing only with the kabyle language. We launched Common Voice along with Mozilla and we started to gather sentences and voices (please see https://voice.mozilla.org/kab), but we need more corpora to launch other processings (morpho-syntactic, grammar, semantics, sentiments analysis .... and so on), thats why we want to launch our locale on tatoeba to gather some specific corpora for these comming NLP tools, the tools are Python-based scripts, and available for free on Gitlab: https://gitlab.com/belkacem77/KabyleNLP , a kabyle morpho-syntactic corpus is also available for free based on a set of 48 tags .

I'd like if you give us a permission to reuse the kabyle corpora you gathered here on tatoeba but also use them on Common Voice and add them for our next corpora on tatoeba. What do you think?

Thanks for your message

M. Belkacem
L10N kab Localizer
Admin of the kabyle locale (Mozilla, Evernote, LibreOffice, VK)