Thursday, April 09, 2009

ANLoc workshop, Ain Sokhna, Egypt

I had the honor of participating in the recent workshop of the African Network for Localisation (ANLoc) in Ain Sukhna (or Sokhna, perhaps a colloquial pronunciation?), Egypt from 31 March to 2 April 2009. The actual location there was the Mövenpick Hotel.

ANLoc is a project including several subprojects and a network of the organizations involved in those efforts, and in localization in general. In effect it is addressing several pieces of what might rightly be regarded as the "foundation" for more effective localization of software and content in African languages. These include:
  • Fonts - Many African languages are written in Latin-based orthographies that include extended characters (modified forms of other letters, often part of the IPA alphabet) and or diacritics that are not found or rare in European languages. Even though all (or almost all) are encoded in Unicode, not many fonts include them yet. This subproject aims to extend some open-source fonts to include such characters.

  • Keyboards - Input of characters not provided for in traditional keyboard layouts requires conceptualizing and designing layouts to accommodate such needs. This is more complex than simple assignment of new characters in a keyboard driver, since issues of anticipating user needs and the possible standardization across languages within a country and across different orthographies in different countries are involved.

  • Language and ICT Policy - There are two major areas of policy (especially of African governments, but also of other entities like development organizations and even industry) that impact the potential for localization - language and ICT policies. Although many do not agree with me, I see the overlap of these two policy areas as an emerging category of policy - "localization (L10n) policy."

  • Locales - A locale (or locale data for a language and geographic location, generally a country) help computers accommodate different language profiles on computers and facilitate localization of software. Developing locales is a key to full presence of African languages in ICT.

  • Localisation Tools - Localization of a software requires expertise in the software and knowledge of the language in which it is to be localized. Not many people have skills in both areas, and even assembling teams that group such expertise is sometimes difficult. Localization tools are really softwares to make the technical aspect of localization easy enough that people with only the requisite language skills can do it.

  • Localise software - This is the object of a lot of the above, and the ANLoc project is sponsoring some pilot initiatives for actual open-source software localization. (One imagines that a future phase of ANLoc might take this on more fully, once the "foundation" work is more complete.)

  • Spell checkers - A key to production of text in any language these days is good spell checkers. Although it might seem straightforward, spell checkers sometimes require a lot of work to accommodate the particularities of how a language may generate words in different contexts.

  • Terminology - How often have you heard in discussion about an African language and technology of any sort that there is no way to say such-and-such in the language? There is a certain truth to the assertion, but it obscures the fact that terms for anything in recent technologies comes from something else or is outright invented. Carrying this process over to several African languages is the object of this subproject.

  • Training - ANLoc will cosponsor some trainings related to localization.

  • Network activities - As a complex project of sub-projects (which together form a network), ANLoc will hold verious meetings.

The workshop in Ain Sokhna provided the opportunity to discuss these various efforts, coordination among them, and suggestions by participants.