Lingua Libre/2025/Phase 1

Constraintes

• 2024 budget:
  • Initial : 18,000€ (DGLFLF)
  • Spent: 2,435€
  • Available: 15,565€.
  • Possible: 10,000€ (WMFR)
• 2025 budget : Unknown.
• Conditions minimales des freelances à WMFR :
  • Par année budgetaire et par prestataire.
  • >0k€ : 1 devis itemisé
  • >2k€ : 3 devis itémisés comparables
  • >5k€ : appel d'offre public
  • >10k€: pre-validation du budget par le Conseil d'Administration de WMFR.
• Par: Prestataire ? Projet ?
• Adhésion : Remy, Xavier
• A summary note is needed.

Introduction

Our fundamental goal is to democratize user-friendly voice and word recording for all languages. We aim to document, preserve, and even revitalize linguistic diversity. This is done within Wikimedia, aligned with its humanistic and free values.

Our target users include all generations, the largest and the smallest communities. These has deep design consequences : Lingua Libre requires broad « ease of use ».

Lingua Libre in 2024

2024 was an important year for LinguaLibre.

On the technical side, Poslovitch lead the creation of a core Lingua Libre Django prototype, Yug lead a double Google Summer of Code 2024 mertorship, onboarding in two Indian developers : Pushkar on Lingua Libre Django and Kabir on SignIt. Two smaller web applications were create. The departure of Poslovitch —now focusing on new projects— presents a major risk for Lingua Libre. Finishing the deployment of Lingua Libre Django and consolidating the developer team are critical.

Community wise, 2024 enabled notable innovative projects and their widespread communication efforts.

• WikiTutur : Ardzun successfully raised two Grants:Rapid, lead dozen of workshops adding 35 new Indonesian languages.
• Youth voices of Roussillon : librarian Culex invited school children to contribute definitions and recordings to wiktionary and Lingualibre.
• Whistled Occitan : Prof. Philippe Biu & Yug lead a documentation and digital revitalization effort for whistled Occitan
• Sign languages : Yug, Brigitte Dale & Kabir (GSoC24) leads French Sign Language and SignIt efforts.

Yug, Ardzun, Culex and others lead over 30 outreach actions, mostly presentations and workshops.

Awareness among both Wikimedians and languages workers have been expanding. For the first time, we saw Lingua Libre groups, projects and supports pop up in East Asia and Africa.

Lingua Libre in 2025

2025 is expected to be a pivotal year for LinguaLibre.org. Web development of the Lingua Libre Django continues. Following users requests to facilitate workshops and diversified contributions, some impactful features are introduced.

WMFR is driving the core of this overhaul through various freelancers.

Certain axis requires wiki expertise from Wikimedia/LinguaLibre contributors. Notably, the migration from the current wiki to Commons, updating User:Lingua Libre bot, promoting it to new wikis, spreading the good news, etc. If you would like to participate in Lingua Libre's 2025 overhaul, contact me, and we can discuss which aspect to tackle.

Current trend and deployment of Lingua Libre Django with easier sharing features should broaden Lingua Libre usages.

Minimal steps to deployment

More front work needed

• T359797: Lingua Libre Django
  • T361440: [GSOC 2024] Pushkar
  • T380121: Milestone 2025Q1 > Freelance : Django #1: split app vs website, i18n, dynamic types, other patch and refresh documentation
  • T384903: Milestone 2025Q1 > Volunteers: review and close tasks
  • Txxxxxx : Milestone 2015Q1.B : App properly connects to Wikimedia Commons API for Oauth and uploads

Database migration is needed

• Properly migrate audio data from LinguaLibre.org's wikibase toward Commons Wikibase
• Properly migrate speakers' data such as learning place, language proeficiency, wikimedia username to local MariaDB database.
• Languages items are already on Wikidata.

Necessary steps for mid terms

• Wikipages migration : migrations of wikipages to Commons. Bulk migration is possible but needs approval within Commons Wikimedia.
• Lists migration : migration within a subdomain of Commons.
• Bot: Revamp User:Lingua Libre Bot

Healthy steps

• Landing website : isolate the static website.
• Doing...Update documentation ./doc
• Recreate statistics dashboard
• Create basic bilingual dictionary viewvers
• Update unit tests

Levers for 2025

Interns

See also Lingua Libre/GSOC25.

The GSoC24 brought in two junior developers, Pushkar (10 weeks) and Kabir (6 weeks), as well as new developers (SalthyKheera).

• Intern: GSoC2025: Stanford / LPI / Tatoeba <-- 4000
• Intern: Outreachy <-- 7000$
• Intern: Service civique <--

Wikimedia collaboration on R&D

• WikiSpeech & AI

Outreach

Reaching out to various communities would have positive impact on the project. Most notably :

• Research partnerships
• Langues de France : previous work by Mélody and other contact list could be leveraged by a service civique.
• Monthly training sessions for Wikimedians and language revitalisation communities

Collaborations

A few ideas and possible fundings :

• Grant:Rapid : Workshops on Kikuyu language, called for by Ngangaesther, National Library Service of Kenya.
• Grant:Rapid : Workshops on Whistled Spanish (Silbo Gomero), Canaries island, called for by Yug.