Template talk:Cite by URL

Conversion to Cargo based implementation

@dcljr: I propose the following transition steps:

1. Convert this into a Cargo based solution, so define the cites in a different page (e.g. Common citations) that contain Cargo store template calls to a new "Citations" Cargo table, and convert this template here to query that table. That would be a 1:1 replacement with no added functionality yet.
2. Modify the various {{Cite …}} templates so that they also store the given data in the "Citations" table. This will likely lead to some duplicate entries, which should not a be problem. We could use queries to detect those duplicate uses and possibly later replace secondary calls with {{Cite by URL}} invocations.
3. Optionally modify the various {{Cite …}} templates so that they act themselves as {{Cite by URL}} calls, if the only data provided is the URL (and parameters like accessed). This is to make adding of citations slightly more comfortable, as it could be tested first if the URL is known by only providing it and generate the preview, and if it isn't known, the missing parameters could be added without the need to rewrite anything.
4. Optionally move the citation definitions from the e.g. "Common citations" page to there natural place, eliminating the need for this extra place. That page could be converted to provide a query that actually showed the most often used entries Though this would require to also store the usages in a table, which would be extra effort, but it could be beneficial to have that capability.
5. Modify this template here so that it also generates output for entries of the Performances database. For performance reasons (no pun intended) this could either be tied to providing some sort of type parameter to distinguish which tables are to be queried, or to keep it simple, be instead implemented in a different template.

--Most2dot0 (talk) 09:01, 7 May 2026 (UTC)

I now did steps 1 and the first part of 2 together, so the common citations page could simply use the regular cite templates. This seems to by working, the only issue I noticed is that an unknown URL will give a "No results" answer instead of a red error note. I assume this is not critical.
Also, there is likely some more thought required how exactly to store and pass the URLs (currently they usually have stored either a leading "//" or "https://", but are accessed without these. That "//" should probably should be included in the WHERE clause (e.i. "WHERE URL LIKE '%//{{{1}}}'" to avoid partial domain string matches, if we leave it that way).
Further, currently the accessed and local values are stored and will be used if available and no new values are given to {{Cite by URL}}. What do you think about that?
In regards to cleanup, I also noticed that in cases were pages are transcluded, like the years articles into the timeline, the entries show up twice, both in the year specific article and for the timeline page. --Most2dot0 (talk) 15:13, 12 May 2026 (UTC)

I updated my last comment by using strikethrough for obsolete items solved in the meantime, and adding some clarifiction in italics. --Most2dot0 (talk) 21:10, 12 May 2026 (UTC)