Software Localization

From Ott09 Wiki
Jump to: navigation, search

Team introductions

Wikimedia foundation, software translators & localization,

Translation vs. localization

  • In localization you have disjoint sentences and words, not "flowing" text
  • Localization frequently refers to tech (mobile phone UIs, software)
  • Subset of translation; mostly same tools but added complexity
    • like placeables
    • if a placeable is noun vs. verb -> differences in translation
    • sentences being broken up
    • space constraints (translation display length)

Strategy; why localize

  • Need to have a strategy
    • Why translate Linux if only 1% uses it?
    • "My mom" wants to use it makes more sense as a goal
  • Tried to come up with "rules"
  • Keep end-user focus
    • Don't translate a programming environment
    • Translate an e-mail program, web browser
  • Emphasize on cross-platform (more value)
    • Like Firefox
  • Translate open source
    • Easier to do
    • Availability of work (whether it succeeds or fails)
  • Helps increase penetration of ICT in general populace (not strategic goal necessary but a good side effect)

Lessons learnt

  • Need to scale the community; can't translate everything
  • Need to prioritize needs
    • Translate the top 25%
    • Easier to achieve quality for this compared to the 100%
  • Translators may do a poor work
    • Example with non-latin characters; needed to purchase a special keyboard
  • Built spell checkers
  • Built tools to help people translate if necessary
    • Pootle (translate online, view progress)
  • small groups (5-7) do a better work at translating
  • translator should be either a user of the software or have some idea what he is translating
  • working now on an offline tool
  • Open source helps create localized environment
    • even as a competition (MS has translated in 5 languages in South Africa)
    • Creating highly-skilled community volunteers is not that easy (they are either not translators or not technically skilled)
  • Attempted to create communities
    • Tried to connect with universities (really low success rate)
    • More successful communities follow the society (wealthier, with better Internet access)
  • Sustainability is very important (since you have to take into account time for software to get a traction and results to be apparent)

Government support

  • Contacted ministers directly (typically not very accessible)
    • Things like meeting someone who knows the minister doesn't work
    • E-mailing the minister doesn't work either; faxing actually worked
  • (Pro) Open source policy of govt helped to a certain degree
  • Education doesn't pay
  • Quite a long term effort; doesn't help if you need help in two months
    • In South Africa it's 8 years down the road after FOSS policy put in place that it gains some traction
  • Required but should not be taken for granted
  • ``Sell`` the fact that software is in local language rather than its Open source nature and merits

Nepal lessons

  • Authored a book on localization (GLOSS; available at
  • Govt assisted by providing a glossary

Sri lanka lessons

  • Tried to create a "terminology bank" especially for stuff like OK/Download/Upload/Authenticate etc.
  • Use Wiki-s to co-ordinate the creation of such banks

Collaboration between FOSS tools

  • Need to see more collaboration between tools builder
  • Goal is to collaborate but to improve as well
  • Hard to do


  • Makes sense to translate just UI and write manual instead of translating help files
  • Workflow: Pootle facilitates a very basic one. Looking into expanding it, so that workflow & translations are on the same file.
  • Politics play a role