User talk:Daniel Mietchen

On this page, old discussions are archived after 90 days. An overview of all archives can be found at this page's archive index. The current archive is located at 2024/08.

SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 1 day and sections whose oldest comment is older than 90 days.

Hello, these items that your bot claimed were published on "19 July 2019" actually are from the XX century!

I suppose this is due to a general misbehavior of the bot. Are you able to identify the magnitude of the problem? And could you fix it as soon as possible? Horcrux (talk) 18:25, 8 July 2024 (UTC)[reply]

@Horcrux: Thanks for checking. The problem is described here, along with a code fix and some discussion about fixing the data. --Daniel Mietchen (talk) 23:12, 9 July 2024 (UTC)[reply]

Yeah, I don't think I'll go into it further. Anyway, some of those items were duplicated of other ones that, curiously, were correctly dated by Research Bot in 2017 (see e.g. [1][2][3]), apart from the known problem of the date's precision ("1" should not be used as day of the month). --Horcrux (talk) 06:59, 10 July 2024 (UTC)[reply]

Hi there, thanks for User:Research Bot which seems very useful and like it has imported data on a significant number of studies.

Could the bot be used to bulk-import data on very many studies – please give your input here and/or here. The latter is about Scholia and I think that platform becomes useful in the real world only once Wikidata has data on, basically, most studies.
For example, charts on scholia would be inaccurate if so many studies are missing in Wikidata. The former is also about books and names a few datasets that I think could be used, maybe most notably dumps by Anna's Archive. I also asked the operator of "LargeDatasetBot". I think you know a lot in this space and your input would be very appreciated!

Also <sub> does not work in item titles – for example see Q125340665. Maybe these could be replaced with the ₂ characters retrospectively and the title be altered accordingly for future imports. Prototyperspective (talk) 18:27, 17 July 2024 (UTC)[reply]

@Prototyperspective: Thanks for thinking along. In terms of imports, the first question should always be whether we should do it, and in the affirmative case, we can explore whether and how we could implement any technical solutions. At the moment, anything large-scale is essentially precluded by the looming graph split anyway. In terms of markup in the titles, I think we should implement something, but technical solutions would probably depend as well on how the graph split plays out. --Daniel Mietchen (talk) 21:14, 21 July 2024 (UTC)[reply]

Thank you, I did not know about this. moving the scholarly articles to a separate graph as in the experiment seems like a good idea and better enable what I asked about. Also I think there need to be some measures to prevent covert vandalism which could already be an issue with such a large number of largely unmonitored items, for example by locking all scholarly articles items to be only changed by bots/queries. The same would also be useful for food product items as mentioned in my recent comment here. I don't know which parts are missing when this bot and a few already imported many items and thought datasets like those used by OpenAlex or ScienceOpen could be used. Prototyperspective (talk) 23:08, 22 July 2024 (UTC)[reply]

This batch has a large number of descriptions that do not follow the guidelines for descriptions. I have initiated a revert of this batch. William Graham (talk) 23:29, 17 August 2024 (UTC)[reply]

Noted. --Daniel Mietchen (talk) 23:50, 17 August 2024 (UTC)[reply]

This is a very bad idea, e.g. https://www.wikidata.org/w/index.php?title=Q35094528&diff=prev&oldid=1616318412. Better not to have a language of work or name (P407) statement than English when it should be French. Charles Matthews (talk) 07:09, 20 August 2024 (UTC)[reply]