Page MenuHomePhabricator

Language Support at Wikimania Hackathon 2024
Closed, ResolvedPublic

Description

The Language Support track will focus on addressing small, language-specific technical requests across various Wikimedia projects to support new and existing language communities. Some of the requests will involve adding and removing languages from projects (e.g., MediaWiki core), adding font and keyboard support for languages, Translatewiki, and other miscellaneous language-related configuration changes such as RTL fixes, and spelling and grammar adjustments.

Tasks for Wikimania will be curated from the following workboards: I18n, WikimediaMessages, MediaWiki-Internationalization, RTL, MediaWiki-General, translatewiki.net, UniversalLanguageSelector, and UniversalLanguageSelector.

Ideas for Tasks

**Curated tasks are under the "Hacking Projects" column (look for tasks with the LPL Technical Support tag): https://phabricator.wikimedia.org/project/view/7284/. These tasks are also added as subtasks to this task and listed below:

Couldn't find anything in this list? Check out the proposed recommendations for the future of language incubation here: https://www.mediawiki.org/wiki/Future_of_Language_Incubation/Recommendations#Proposed_Recommendations

Sessions

As part of the Language Support track, there will be a:

  • Help desk to allow participants to chat with track organizers about language tools, ask technical questions, and share feedback on the tools they use.
  • Technical support provided to interested volunteers working on language-specific technical tasks on the Technical Support Workboard. Track leads will share task ideas, onboard volunteers (finding code repositories, providing technical guidance, assisting with code reviews, and merging).
  • Technical sessions related to the track to provide an overview of the language support area, language tools and projects, volunteering opportunities, and more.

Track organizers: @srishakatux, @Arrbee, @UOzurumba, @Aaharoni-WMF, @jhsoby

Related Objects

Event Timeline

debt updated the task description. (Show Details)
Amire80 renamed this task from Language Support to Language Support at Wikimedia Hackathon 2024.Jul 7 2024, 5:23 PM
Aklapper renamed this task from Language Support at Wikimedia Hackathon 2024 to Language Support at Wikimania Hackathon 2024.Jul 7 2024, 10:04 PM

@srishakatux, @MaryMunyoki: Hi,

  • If this initiative took place: Please change the task status to resolved via the Add Action...Change Status dropdown.
  • In this session / event did not take place: Please set the task status to declined.

Thank you, Phabricator housekeeping service

closing this ticket as resolved - Wikimania is over, thanks to all!

Notes from Language Support Activities

Some work from this track includes:

  • Adding RTL support for the Internet Archive Bot (T372035).
  • Adding namespace translations for the two South African languages, Xhosa and Swazi (T372057).
  • Adding the languages Duala (T372173) and Ngiemboon (T371985) from Cameroon to Translatewiki.net.
  • Adding the Silesian alphabet to jquery.ime (T372652).
  • Making the Sarjom baha layout the default for Santali Wikipedia (T372146).
  • Hosting a session on translatewiki.net (T371535) led by Uzoma for remote attendees.
  • Other noteworthy projects related to this track that were worked on by others include:
    • The Toolforge i18n library (based on the Wikidata Lexeme Forms tool) and related documentation for using it with Python tools (T283376).
    • Adding logos for newly created Malaysian wikis, including the Kadazandusun Wikipedia (T372031).

During the Hackathon, there were dedicated tables for each product and tech initiative area that were the focus of the event, including language support. Several community members shared their feedback on language tools. Although we directed many of these individuals to existing resources and our usual channels for sharing requests, I’m documenting some of them below for the record:

  • Dark mode is missing in Content Translation.
  • With Translatable pages, messages in the content break when they are inside a notice tag.
  • Section translation is missing on Vietnamese Wikipedia.
  • The Content Translation tool doesn’t detect reverts if the content of the original article changes. The publish button gets disabled if changes occur in the original article.
  • Integration of the DeepL translate tool with Content Translation.
  • On the desktop version on mobile, there is a translations tab that takes you to the Content Translation tool. In the mobile version on mobile, there is no way to access the translations tab. A link should be added.
  • Machine translation quality is poor on Russian Wikipedia.

In conversations with several members of newer and smaller language wikis, such as Kadazandusun (Malaysia), Santali (India), and Dagbani (Ghana), a few common challenges emerged:

  • When growing in Incubator, it is often the more tech-savvy editors who perform complex operations on the wiki, such as copy-pasting, using and writing templates, and importing content with tools such as QuickStatements. Communities provide training to their contributors to become comfortable using wikitext and performing more complex tasks. One person mentioned, “If we can reduce the use of templates for smaller wikis, that might be helpful; it is difficult for them as it requires technical knowledge.” Another said, “If we encounter any errors while using templates, we don’t investigate them.”
  • Communities almost always have no one to set up bots for them. One contributor mentioned they tried to set up a welcome bot, but it was hard, and they eventually gave up.
  • Kadazandusun community members mentioned there is no machine translation available for their language. We wondered if there are language models for low-resource Southeast Asian languages or if the Teacher Institute at Kent University, which is heavily focused on reviving the language, could collaborate with others on developing these models by learning from experiences such as the IndicTrans project.
  • It was a personal observation from many of the conversations that bug reporting venues in Wikimedia are well-known to those who are part of our community, are developers, or are familiar with reporting bugs, using Phabricator, or following the usual channels. For others, the process is not as apparent, and the available resources, such as the footer section on any Wikipedia page with a "Developer" link leading to Developer resources, are not very inviting to our users.

@PDas, @SGill, and I met to discuss engagement opportunities for LPL Technical Support with WikiTechClub (India). We explored the possibility of including language and internationalization in software projects as one of the topics for their curriculum, which focuses on Wikimedia projects and technical areas. We also discussed providing a monthly sync meeting and more asynchronous support, offering tasks and guidance to potential candidates.

Thank you to @jhsoby, @UOzurumba, and @Amire80 for all their work, and to everyone else who contributed to this track!

Notes from Language Support Activities

Some work from this track includes:

  • Adding RTL support for the Internet Archive Bot (T372035).
  • Adding namespace translations for the two South African languages, Xhosa and Swazi (T372057).
  • Adding the languages Duala (T372173) and Ngiemboon (T371985) from Cameroon to Translatewiki.net.
  • Adding the Silesian alphabet to jquery.ime (T372652).
  • Making the Sarjom baha layout the default for Santali Wikipedia (T372146).
  • Hosting a session on translatewiki.net (T371535) led by Uzoma for remote attendees.
  • Other noteworthy projects related to this track that were worked on by others include:
    • The Toolforge i18n library (based on the Wikidata Lexeme Forms tool) and related documentation for using it with Python tools (T283376).
    • Adding logos for newly created Malaysian wikis, including the Kadazandusun Wikipedia (T372031).

During the Hackathon, there were dedicated tables for each product and tech initiative area that were the focus of the event, including language support. Several community members shared their feedback on language tools. Although we directed many of these individuals to existing resources and our usual channels for sharing requests, I’m documenting some of them below for the record:

  • Dark mode is missing in Content Translation.
  • With Translatable pages, messages in the content break when they are inside a notice tag.
  • Section translation is missing on Vietnamese Wikipedia.
  • The Content Translation tool doesn’t detect reverts if the content of the original article changes. The publish button gets disabled if changes occur in the original article.
  • Integration of the DeepL translate tool with Content Translation.
  • On the desktop version on mobile, there is a translations tab that takes you to the Content Translation tool. In the mobile version on mobile, there is no way to access the translations tab. A link should be added.
  • Machine translation quality is poor on Russian Wikipedia.

In conversations with several members of newer and smaller language wikis, such as Kadazandusun (Malaysia), Santali (India), and Dagbani (Ghana), a few common challenges emerged:

  • When growing in Incubator, it is often the more tech-savvy editors who perform complex operations on the wiki, such as copy-pasting, using and writing templates, and importing content with tools such as QuickStatements. Communities provide training to their contributors to become comfortable using wikitext and performing more complex tasks. One person mentioned, “If we can reduce the use of templates for smaller wikis, that might be helpful; it is difficult for them as it requires technical knowledge.” Another said, “If we encounter any errors while using templates, we don’t investigate them.”
  • Communities almost always have no one to set up bots for them. One contributor mentioned they tried to set up a welcome bot, but it was hard, and they eventually gave up.
  • Kadazandusun community members mentioned there is no machine translation available for their language. We wondered if there are language models for low-resource Southeast Asian languages or if the Teacher Institute at Kent University, which is heavily focused on reviving the language, could collaborate with others on developing these models by learning from experiences such as the IndicTrans project.
  • It was a personal observation from many of the conversations that bug reporting venues in Wikimedia are well-known to those who are part of our community, are developers, or are familiar with reporting bugs, using Phabricator, or following the usual channels. For others, the process is not as apparent, and the available resources, such as the footer section on any Wikipedia page with a "Developer" link leading to Developer resources, are not very inviting to our users.

@PDas, @SGill, and I met to discuss engagement opportunities for LPL Technical Support with WikiTechClub (India). We explored the possibility of including language and internationalization in software projects as one of the topics for their curriculum, which focuses on Wikimedia projects and technical areas. We also discussed providing a monthly sync meeting and more asynchronous support, offering tasks and guidance to potential candidates.

Thank you to @jhsoby, @UOzurumba, and @Amire80 for all their work, and to everyone else who contributed to this track!