You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I don't remember if we discussed this before specifically about CitationHunt, but how hard would it be in your opinion to add a feature which actually suggests possible citations rather than just report the lack thereof?
I believe the Internet Archive already has tools which would allow to search sources for a specific information across millions of books, some 18M papers and many more millions of other items they host. I'm told by Mark G. of IA that he has some prototype already, no idea how hard to integrate.
Hey, thanks for the idea! Would you happen to have some documentation or more information about that prototype? By the way, I'll be at Wikimania starting on Friday, we can chat if either you or Mark are around :)
My main concern with adding this feature is the accuracy of the recommender, as I haven't seen a particularly impressive one yet, but I have seen some research in that direction at WikiCite.
That is, if we're "just" searching by the text of the snippet on a large corpus of documents, and returning a few documents that seem to match the snippet for the user to triage, then we really need to ask ourselves whether the feature is providing more value than a regular search engine that the users would have already gone to on their own.
But yes, if there is such a service that provides reasonably localized suggestions given a snippet, integrating it in CH would be the easy part. If I can help build or validate that service, by the way, I might also be able to find some time!
I don't remember if we discussed this before specifically about CitationHunt, but how hard would it be in your opinion to add a feature which actually suggests possible citations rather than just report the lack thereof?
I believe the Internet Archive already has tools which would allow to search sources for a specific information across millions of books, some 18M papers and many more millions of other items they host. I'm told by Mark G. of IA that he has some prototype already, no idea how hard to integrate.
Sort of related efforts are
https://www.wikidata.org/wiki/Wikidata:Primary_sources_tool
https://meta.wikimedia.org/wiki/Research:Expanding_Wikipedia_articles_across_languages/Intra_language_approach
https://meta.wikimedia.org/wiki/Grants:Project/Frimelle_and_Hadyelsahar/Scribe:_Supporting_Under-resourced_Wikipedia_Editors_in_Creating_New_Articles
The text was updated successfully, but these errors were encountered: