[13:29:42] Forwarded from relgu: For those of you who are interested in seeing how a templatic NLG solution for Abstract Wikipedia, using Wikidata lexemes, could work in practice, I have uploaded to Wikicommons a demo video of my Scribunto-based prototype of such a system: https://meta.wikimedia.org/wiki/File:Demo_of_a_Scribunto-based_templatic_NLG_system.webm [13:59:16] Thanks for the share relgu! I am currently looking at it, it's interesting [16:06:44] That's a fantastic demo! I must say that on the one hand I find it exciting and reassuring, by offering a more concrete and somewhat complete example of what Abstract Wikipedia content could be assembled and how could look like (compared to what I had seen so far), but on the other hand I find it concerning, in that the whole system is rather complex and not something I could see [16:06:45] thousands (hopefully millions) of volunteers joining around to create a community of editors, unlike lower-hanging fruit like Wikipedia articles, media in Commons, and to some extent basic data statements in Wikidata. (re @relgu: For those of you who are interested in seeing how a templatic NLG solution for Abstract Wikipedia, using Wikidata lexemes, could...) [16:10:53] That said, it is true that we do have precedent in, say, Wikimedia templates/modules (that are built by a highly specialized subset of the community), and in developers of open source software, etc. It is arguable that a gradual disclosure of complexity may have helped in those cases (e.g. one can build very simple templates that are already useful), which I'm not sure is possibl [16:10:54] e with NLG. But as long as there's a wider community of Wikimedians, and given the large payoff of this work, I suppose a smaller subset of "Abstract Wikipedians" could indeed thrive, unlike, say, obscure WikiProjects that end up becoming dormant after a while. Anyway, exciting times! [16:48:51] Thanks for the positive feedback! [16:48:51] It is true that the system in its integrality is complex, but the hope is that for every language, only a handful of contributors will have to deal with the low-level complex functions, and most of the contributors could write templates for specific constructors, which should be easier. [22:53:39] Saluton [22:58:15] /search@wikilinksbot [22:58:36] Sorry