Community Wishlist Survey 2022/Wikidata/Multilingual dictionary GUI using Wikidata lexemes

Multilingual dictionary GUI using Wikidata lexemes

  • Problem: Wikidata contains lexical items that can be displayed as translation equivalents for a word sense: (1) lexeme senses in two languages that are linked to the same item with property P5137 'item for this sense', (2) lexeme senses of one of the language that are linked to some item with property P5137, and that item having a label in the other language, (3) item labels, if an item has labels in the desired languages, (4) repeating the same with synonyms (P5973), (5) P5972 sense translation links. It would be helpful to have a tool like Hauki (the proposed tool could well be based on that), which lets select desired source language and target language(s), and displays all found equivalents, combining the above queries, and that lists the results, referencing the source.
  • Proposed solution: A tool e.g. based on Hauki that displays translation equivalents from the different sources described. Two kinds of use cases:
  1. Translation equivalent lookup. A view like in a dictionary entry would be helpful, with part-of-speech sections and indented sense blocks, and some visual markup of different equivalent types (P5137-linked sense lexicalisations, P5973 synonyms, any P9970 'predicate for' relation, labels and aliases of the pivot item), and, if available, sense glosses, examples, etc.
  2. P5137/P5972/P5973-editing task. Propose item through reconciling a sense lexicalisation against item labels, and also give linked multilingual lexicalisations and glosses as support for the manual validation decision of the user. Ideal would be one-clik validation or rejection of the proposed P5137/P5972/P5973-links. This would shortcut going through OpenRefine, and additionally offers multilingual info, already linked to the proposed item. That is very useful for multilingual users.
  • Who would benefit:
    • Users who want to find translation equivalents
    • Lexicographers
    • Wikidata contributors, or the task of adding missing P5137/P5972/P5973 statements
  • More comments: Lexicographical data on Wikidata is becoming more and more powerful; P5137/P5972/P5973 coverage is a key issue for further improvement towards a use for a human-readable multilingual dictionary.
  • Phabricator tickets:
  • Proposer: DL2204 (talk) 19:18, 22 January 2022 (UTC)

Discussion

Voting