Wikimedia Language engineering/Reports/2017/May

April 2017

Wikimedia Language engineering monthly report for May 2017

June 2017

Monthly updates from the Wikimedia Language team focused on technical development of the software they maintain. Names of people assumed to contribute as volunteers are in bold text.

Highlights

• During the Vienna Hackathon, Siebrand Mazeland, Federico Leva and Niklas Laxström processed many support requests in the backlog, resulting for example in a dozens new languages being enabled in translatewiki.net
• Translatewiki.net had an outage that lasted 25 hours. A retrospective held in Vienna Hackathon identified multiple action points for improvements.
• Amir Aharoni's Telegram bot was worked during the Vienna Hackathon. It allows to translate messages with your phone using Telegram.
• MediaWiki now supports Kabiyè language.
• Special:LanguageStats is now more reliable and faster, with more speed improvements planned.

Content Translation

Content Translation allows people to translate Wikipedia articles. Also check out out the newsletter.

About 40 patches were merged this month.

• Santhosh Thottingal improved link handling in rewritten Content Translation.
• Ed Sanders updated code style checker tool.
• Niklas Laxström integrated machine translation services to rewritten Content Translation. T162110
• Santhosh Thottingal updated code to avoid using code deprecated in the OOJS UI library. T155152
• Santhosh Thottingal added initial code architecture documentation.
• Santhosh Thottingal updated code to avoid using deprecated icon packs in the OOJS UI library. T166730
• Santhosh Thottingal added support for easy namespace selection for the title to publish. T165562

Usage data

Details	Numbers
Articles created (during the month)	10559
Articles created till end of month	206383
Number of new translators (during the month)	1070
Highest number of articles created by one user (cumulative number till end of the month)	4517
Highest number of translators for a Wikipedia (cumulative number till end of the month)	4806 (es)
Total number of translators with only 1 published article (cumulative number till end of the month)	18859
Total number of translators with only 2 published articles (cumulative number till end of the month)	3879
Total number of translators with only 3 published articles (cumulative number till end of the month)	1826
Total number of translators with only 4 published articles (cumulative number till end of the month)	1037
Total number of translators with >= 5 and < 20 published articles (cumulative number till end of the month)	3018
Total number of translators with >= 20 and < 100 published articles (cumulative number till end of the month)	1086
Total number of translators with >= 100 published articles (cumulative number till end of the month)	305
Articles deleted (during the month)	995

Note: A spike in deletions caused by deletion of about 400 articles created as a vandalism in Kazakh Wikipedia.

Comparisons

• 
  Articles published.
• 
  Deletion ratio.
• 
  Distribution of users per number of translations they published.

Translate

Translate extension allows translation of software interface, manuals, announcements, banners etc. on Wikimedia special projects, translatewiki.net and other sites. Translate is part of MLEB.

About 20 patches were merged in this month.

• Niklas Laxström optimized a piece of code and made Special:LanguageStats much faster when it contains many aggregate message groups. T160565
• Niklas Laxström fixed an issue that caused newlines in translations to appear incorrectly in the Android localisation file format.
• Ed Sanders updated JavaScript coding style checker and made the code to pass the new rules.
• Aaron Schulz made more changes to Translate to avoid querying the master database server unnecessarily.
• Niklas Laxström fixed a broken error propagation in translator sandbox user creation in case of an invalid password. T164912
• Niklas Laxström added experimental support for Caighdean machine translation service.
• Niklas Laxström improved error handling for web services providing translation suggestions.
• Niklas Laxström rewrote the background job that is used for renaming translatable pages.
• User TerraCodes removed redundant dialog title from the message group selector. T165241
• Niklas Laxström made Translate extension easier to identify by always adding User-Agent header when doing requests for web services providing translation suggestions.
• Sam Reed removed the bundled Spyc library. Now it must be installed via composer if needed. HHVM users don't need it, nor any users not using any YAML files for configuration or translation. T75945
• Amir Sarabadani and Thiemo Mättig improved the code generating colors in Special:LanguageStats and Special:MessageGroupStats. T162058
• Niklas Laxström fixed CXServer web service that was broken for a month or two and only showing empty suggestions. T165949

Usage data

• Translatable pages: Meta 3708 (+76); MediaWiki.org 3405 (+27); Commons 633 (+9) [1][2][3]
• Translate extension is installed on 215 wikis (+5) Wikiapiary

Universal Language Selector

Universal Language Selector provides language choice, input methods and web fonts. It is part of MLEB.

• Amir Aharoni updated upstream libraries jquery.uls and jquery.ime.
• Kartik Mistry corrected a comment which was using an outdated name of a script file.
• User Fomafix added a document type declaration to generated HTML file.
• James Forrester and Timo Tijhof removed unconditionally dependency on a deprecated JavaScript module in MediaWiki core. T162590
• Ed Sanders updated JavaScript coding style checker.
• User Isarra improved the callout dialog positioning for different skins. T161586

Other extensions

Babel (part of MLEB):

• Amir Aharoni added caching for fetching user language skills, so that it can be used in more places. T164187
• User TTO improved handling of MediaWiki's non-standard language codes. T101086

cldr (Common Locale Data Repository) (part of MLEB):

• Ed Sanders replaced existing JavaScript coding style checker with eslint.
• User Pikne added Estonian translators for language names not present in CLDR.

CleanChanges (part of MLEB):

• No changes.

LocalisationUpdate (part of MLEB):

• No changes.

TranslationNotifications:

• No changes.

MLEB

• No MLEB was released in this month. Next release is planned for July.

MediaWiki core internationalization

Changes to translations and equivalent localisation data (magic words, namespace names) are not included in this list. Huge thanks are given to our volunteer translators as well.

• Derk-Jan Hartman and Matthias Mullie fixed an issue that caused proper language and directionality attributes to be missing from redirect pages in the file namespace. T60955
• Brad Jorsch made cleanups to MessageCache. T164666
• Brian Wolff added collation for Bashkir language. T162823
• Brad Jorsch made it possible to localise all errors in the API login module.
• Sébastien Santoro changed Nepalese to use Devanagari digits in HTML lists. T151633
• User Chiefwei updated language converter data for Chinese.
• Justin Du cleaned up code in MessageCache.
• Thiemo Mättig updated documentation for Language::getDatePreferences.
• MediaWiki now supports Kabiyè language. T160915
• Brian Wolff fixed an issue that broke collation in Persian language. T139110
• Kunal Mehta cleaned up code in LanguageConverter.

Project Milkshake

Project Milkshake consists of four re-usable jquery internationalization libraries.

jquery.i18n

• Santhosh Thottingal fixed an error when trying to load fallback languages.

jquery.uls

• Amir Aharoni added five language of Taiwan to the language database.
• Kartik Mistry updated bundled Spyc library.
• Amir Aharoni added Rangi language to the language database.
• Federico Leva added four new languages for translatewiki.net to the language database.
• Federico Leva added variants for Hakka language to the language database.

jquery.webfonts

• No changes.

jquery.ime

• Amir Aharoni added input methods for Dinka, Bambara and Dagbani languages.
• Amir Aharoni fixed right-to-left bugs.

translatewiki.net

See also translatewiki.net news page.

• Raimond Spekking added new MediaWiki extensions and skins for translation, made changes to many and did exports and imports for all MediaWiki.
• Niklas Laxström updated SSL configuration.
• Amir Aharoni added many new languages for translation.
• Federico Leva updated settings for Potlatch.
• FUEL and Europeana are no longer available for translation.
• Niklas Laxström updated HHVM configuration.
• Niklas Laxström installed OAuth extension to support Amir's Telegram translation bot.
• Niklas Laxström added a custom error page to nginx.
• There was a long, 25 hours long, outage due to issues with hosting. A retrospective was held in the Vienna Hackathon. T165539
• Siebrand Mazeland added many new languages for translation.
• Siebrand Mazeland and Niklas Laxström continued unification of access control for all projects to remove dependency on any single person to do translation exports.
• There was a small outage due to denial of service caused by a bug in LiquidThreads.

Stats

• About 80 (+15) new users
• About 330 (=) translators made about 42.000 (+12.000) edits
• About 90 (-5) translators reviewed about 10.500 (+1.500) translations

Sources: [4] [5] [6] [7] [8]

TwnMainPage

• No changes.

Deployments and other site related updates

• Kartik Mistry added non-default MT support for Burmese, Khmer and Lao languages.

Cross team work/requirements

• No updates.

Team status

• Niklas Laxström attended the Wikimedia Hackathon in Vienna and wrote a report.