LinguaLibre:Events/2022.06 Lingualibre online meetup/de

Das LinguaLibre-Onlinetreffen 2022.06 (1. Juni 2022) ist das Wikimedia-Frankreich-LinguaLibre-Treffen zur Jahresmitte, um die konferenzreiche Sommersaison, das spätere Jahr 2022 und die mittelfristigen Wünsche für 2023 zu planen.

Details

• Beschreibung: Online-Treffen zur Jahresmitte zur Vorbereitung des konferenzreichen Sommers für Veranstaltungen Ende 2022 und 2023.
• Ort: Online
• Uhrzeit/Zeitplan: 2022/06/01 > 20:00 UTC+2.
• Organisation: User:Adélaïde Calais WMFr, Wikimedia Frankreich.
• Teilnehmer: Poslovitch, Lyokoï, Rdrg109, Rodion (LanguageSeeker), 0x010C, Yug, Assassas77.
• Etherpad: https://etherpad.wikimedia.org/p/2022.06_Lingualibre
• Visio: https://visio.wikimedia.fr/aca-ext-jtf-ubi
• Dokumente: meta:User:Adélaïde Calais WMFr/2022-2023 Lingualibre wishlist Tabelle und Abhängigkeitsdiagramm.
• Erhebnis: Teambuilding mit Neulingen, Besprechen kommender Veranstaltungen, meta:User:Adélaïde Calais WMFr/2022-2023 Lingualibre wishlist, Status von Lingualibre und Bedarf an menschlichen Ressourcen bei wichtigen Engpässen.
• Überprüfung & Verbesserungen: Eine aktive Planung sechs Monate im Voraus wäre von Vorteil.

Missions map

1. MediaWiki maintenance and enhancing : solving underlying technical issues (SPARQL, lack of a developper environment, heavy bot, record wizard lacking beginners tools etc.)
2. Outreach : get more speakers, get more records
3. Features developments
4. Fund requests, aka money money money.

See also recent wishlist.

Directional Acyclic Graph of 2022-2023 Lingualibre wishes (meta:User:Adelaide_Calais_WMFr/2022-2023_Lingualibre_wishlist)

Translations

Not discussed.

Calendar

Contextual dates and deadlines

• 2022.03.11: Adelaide initiated the 2022-2023 Lingualibre wishlist.
• 2022.06.01 (event): Updated wishlist shared, call for improvement.
• 2022.06.05 (deadline): meta:Celtic Knot Conference 2022 > submission deadline. A conference for “communities working on a minority language on the Wikimedia projects”, See requested format: <5mn prerecorded video.
• 2022.06.10 (deadline): Wikimania 2022 submition deadline. See possible formats.
• 2022.06.21-23 (event): Lingua libre in Marseille. Opportunity to share our vision with French tech companies.
• 2022.07.01-02 (event): Celtic Knot Conference 2022.
• 2022.07.08-10 (event): Wikimedia France's Wikicamp. Opportunity to share our vision with WMFr's community.
• 2022.08.11-14 (event): Wikimania Online. Opportunity to share our vision with the global WMF's community.

Minutes

This section is messy and unclear. Clarification welcome.

• 20:06 Wikimania - Lyokoï wants to organize a contribution event around Lingua Libre - challenge style. "How many languages can we record on lingua libre in one day ?"
• Idea : need to prepare lists for people to pronounce. Rodio has been working on a bot to generate lists of words that are missing . Olaf bot does that (grabbing words from wiktionnaries).
• Tech days at Wikimania - last time turned into a presentation more than a hackathon. The issues we have are too complicated for newcomers to work on them. Rule out a hackathon.
  • Note for the future : it would be nice to have a list that is more fluid - a tool that doesn't present the same words to every participant.
• 20:21 Poslovitch: The recording process is intimidating for some users.
  • RecordWizard url link with list loading
• 20:27 Lyokoi: It is not a problem that we have multiple recordings of the same word since the accent, the age or gender of each recording is also important
• 20:31 Rodion: At what point can we say that it's enough for the number of times the same word is recorded? 30 recordings? 100 recordings?
  • 20:34 Rodion: When new users are willing to contribute, a list of words could be suggested and those words would be those that don't have much recordings
• [Gap in vision of what Lingua Libre is , and what it is for]
• 20:37 Poslovitch: Theme/topic generated lists was discussed in June, Hackaton 2021 at Lyon.
• New list types to think about: frequency list, never recorded lists, thematic lists.
• 20:44 acalai There will be an event in Massy, south of France, aimed to tech companies and tech researchers who work in languages resources. It lasts for 3 days.

• Events:
  • Yug and Lyokoï - Marseilles

• meta:User:Adélaïde_Calais_WMFr/2022-2023_Lingualibre_wishlist
• 20:50 acalai MediaWiki update in spring 2023. Need to fix compatibilities.
• Click bug:
  • Yug: Bottleneck issues to solves are the sound library and audio click bug.
  • Adelaide: never saw ot, is it frequent ?
  • Yug: occurence rate undocumented.
  • 21:04 Poslovitch: Need Matomo to find those bugs that are affecting some users but haven't been solved yet due to the lack of a description of the bug. Use the proper property on each recording that is faulty. problem : when you've recorded 1000 words, no one wants to look at each individually and manually put the property on each of them. -> would need a batch editing tool (for power users)
    • -> step 1 : create batch editing tool
    • -> step 2 : have a human patrol campaign to create a corpus of faulty recordings
    • -> step 3 :
• 21:13 Poslovitch : lingua libre is no longer only a recording tool, it needs to be a sound library as well.
• Routines
  • The datasets don't work.
  • Lingua Libre bot has not been working for 7~8 weeks.
• (Who?): SPARQL scares new developers (because it's hard and slow). Need json files of the datasets.
  • Yug: Help:SPARQL now available. Authors Yug, VIGNERON, Rdrg109
• IDE:
  • Poslovitch :Needs developer environment - how to set up lingua libre on one's computer.
• Need processes and a yearly hackathon

Outreach for resources

• A developer for 1 year ?
• University partnership - 1 lab of researchers to contribute to + use
• Outreach for money / Fundraising

How to communicate with decision makers:

• To get fund : prove strategic role !
• Strategic arguments : diversity, revitalization, orality, reusability (forvo not very reusable), document and preserve dying languages

Francophonie.

• Ask questions when WMFr board meeting holds public discussion.