Complex translation workflow and formatting

Supporting a multilingual application has been a difficult for a very long time and we are now in a brink of entering a new state where we can do it more easily. Data such as CLDR and technologies such ICU’s MessageFormat is unheard for most people. Yet they are used by global companies such as Google, IBM and Apple. These technologies is now included in a project I have been working on called L10ns(). In addition to solving complex translation formatting with the help of ICU’s MessageFormat and CLDR, L10ns also solves a very difficult workflow problem associated with translation. My session will focus on all the mayor problems in internationalizing applications and help people use efficient industry best practices and solutions and l10ns(http://l10ns.org) for producing multilingual applications.

Transcript

1. Translation workflow and formatting complex translations Tingan Ho @tingan87 L10ns
2. Structure of this talk • Demo of problem and solutions regarding i18n Two parts: • One part explaining complex translation formatting • Other part explaining i18n workflow Publish l10ns
3. Part1: Complex translation formatting • CLDR and ICU’s messageformat • Plural formatting • Context based formatting • Number formatting • Date formatting
4. Pluralization: bad example 1
5. Pluralization
6. Cons: Locale specific code block • Increased programmer dependecy from translator • Maintenance problem - remembering locale specific data
7. So how do we solve this problem?
8. How do we make people not crying over their code base?
9. CLDR (Common Locale Data Repository)
10. CLDR • Data used for localizing strings • Maybe the largest locale data repository in the world • Used by Apple, Google, IBM Yahoo etc.
11. CLDR - example of XML data
12. CLDR - plural forms • Spoken English just have two defined plural forms namely, singular and plural. • CLDR defines up to six plural forms. Zero, one, two, few, many and other
13. Remember it! zero, one, two, few, many, other
14. zero, one, two != 0,1,2 CLDR - plural forms
15. CLDR - plural forms examples • English: one, other. • Arabic: zero, one, two, few, many and other. • Chinese: other • French: one, other in French: plural form one = 0, 1
16. CLDR plural form
17. ICU’s Messageformat
18. General message formatting • String concatination • Template strings
19. ICU’s messageformat • ICU stands for International components for Unicode (written in C++ and Java) • ICU’s messageformat is a component in ICU for handling message strings
20. ICU’s messageformat - plural format
21. Try to define a simpler syntax... using at least… • one variable(counter) • type of formatting • messages for each plural form
22. ICU’s messageformat - plural format {variable, plural, one{...} other{...}}
23. ICU’s messageformat - plural format {variable, plural, zero{...} one{...} two{...} few{...} many{...} other{...}} Arabic
24. Demo (I liked x blog post/s)
25. Plural format
26. Select format
27. Select format use case Context based formatting
28. X likes this conference • He likes this conference • She likes this conference • They likes this conference
29. ICU’s messageformat - select format {variable, select, case1{...} casen {...} other{...}}
30. ICU’s messageformat - select format {gender, select, male{He} female {She} other{They}}
31. Select format
32. Number format
33. ICU’s number format {variable, number, #,##0.00} {variable, number, integer} {variable, number, percent} {variable, number, permille}
34. ICU’s number format {variable, number, ^¤#,##0.00}
35. Number format
36. Date format
37. ICU’s date format {variable, date, yyyy-mm-dd HH:mm:ss}
38. Part2: Translation workflow • Syncing translation keys • Translator and programmer workflow • Translation key naming
39. Syncing translation keys Storage Source code Sync
41. Manual sync • When you delete a key in source you need to delete it from storage. • When you add a key in your source you need to add it in your storage. • When you change a key in your source you need to change it in your storage.
43. Most projects stay out of sync
44. How can we fix it?
45. Auto sync
46. Auto sync • When you delete a key in source it will delete it from the storage • When you add a key in your source it will add it in your storage • When you change a key in your source it will migrate to the new key
47. Auto sync work flow 1. Write your source code 2. Update translation keys 3. Translate 4. Compile your translations 5. Run your application
48. Demo (auto sync addition, deletion, migration)
49. Translator and programmer workflow
50. Translator and programmer workflow 1. Try to always store the English translation, so they can relate to the English translation. 2. Let the translator learn about ICU’s messageformat. 3. Have a web translation interface where they can translate your strings.
51. Translation key naming
52. Anti pattern Don’t use English text as translation keys. Because you will have namespace issue and no future proof.
53. Content structure - facebook Feed FEED__POST_LIKES FEED__POST_COMMENTS Side Menu SIDE_MENU__NEWS_FEED SIDE_MENU__MESSAGES
54. Translation workflow
55. Launch
56. Links L10ns http://l10ns.org/ ICU http://site.icu-project.org/ CLDR http://cldr.unicode.org/
57. Thank you! http://l10ns.org Tingan Ho @tingan87