User talk:John Cummings/Wikidata in Wikimedia projects
Dump of previous discussions
[edit]You may wish to scan these for arguments: there are plenty of them.
- en:Wikipedia:Requests for comment/Wikidata Phase 2
- en:Template talk:Infobox book #Wikidata
- en:Wikipedia:Village pump (policy)/Archive 128 #RfC: Wikidata in infoboxes, opt-in or opt-out?
- en:Wikipedia:Templates for discussion/Log/2017 May 11 #Template:Infobox_person/Wikidata
- en:Wikipedia:Wikidata/2017 State of affairs - and its talkpage
You may also wish to read the documentation for en:Module:Wikidata and en:Module:WikidataIB in light of the strength of feeling about local overrides and opt-in vs opt-out. That might help you with seeing why I introduced the second module specifically to implement whitelists and blacklists in infoboxes, as well as providing filters to allow only sourced statements to be imported into an infobox. --RexxS (talk) 18:20, 24 September 2017 (UTC)
- I'll just add some (older) German discussions:
- de:Wikipedia:Meinungsbilder/Nutzung von Daten aus Wikidata im ANR
- de:Wikipedia:Umfragen/Integration Wikidata in Wikipedia
- de:Wikipedia:Projektdiskussion/Wikidata/Archiv/1 mostly inactive nowadays de:Wikipedia:Projektdiskussion/Wikidata
- de:Wikipedia Diskussion:WikiProjekt Wikidata in Wikipedia
- de:Wikipedia:Projektdiskussion/Wikidata-Beschreibung unter der Artikelüberschrift für die mobile Version
- hope this helps. --HHill (talk) 20:46, 28 September 2017 (UTC)
Defensive programming
[edit]Something to add might be how to use Wikidata in a "defensive" way, perhaps for most visible/most sensitive areas.
For example, moving data completely to Wikidata might be appropriate for particular external links templates. (Though the ability to specify a local override, in local wikitext, can be useful - on both scores).
But for other data, it may make sense to go on carrying a copy locally, but have the template flag if the local copy and the Wikidata copy become different -- eg by setting a hidden template. Examples of best practice using this approach in the field, and code available to support it on particular wikis, could both be useful. Also whether such templates should also have a slot for eg a WONT_SYNC flag, to indicate that such a difference is to be expected, or has been reviewed and accepted. Jheald (talk) 15:39, 5 October 2017 (UTC)
Wikidata for data management at a project level
[edit]Moving away from a focus starting on single articles, Wikidata has a lot to offer projects in terms of consistent and systematic presentation of information.
Queries make it easy to see what information is held for a particular set of articles, and what gaps, anomalies, or variations in approach it may contain.
Quick Statements makes it much easier to standardise or fix things at scale (eg a set of external urls some of which may have been revised), compared to what has to be done to similarly update wikitext at scale. (QS is far less bother than AWB; & a lot easier than having to write/re-write a bot).
A nice application of Listeria reports & "view history" can make it possible to see information relevant to the project that has been changed over eg the last month, or 3 months. This is useful (a) for giving an idea of hot areas of activity; but also (b) for catching up and reviewing "bad edits", that may have got through an initial vandal screen.