Wikipedia talk:Special:UnwatchedPages

This is the talk page for the special page at Special:UnwatchedPages. For general information on this and other special pages, see Help:Special page. For recent talk about special pages, see Recentchangeslinked/Specialpages discussion

Archives

Archive 1

This page has archives. Sections older than 30 days may be automatically archived by Lowercase sigmabot III when more than 4 sections are present.

See Wikipedia:VPT#Number of unwatched pages – Philosopher ^{Let us reason together.} 23:31, 15 December 2014 (UTC)[reply]

Late update: Zero in mainspace. Now let's keep it at that. ~ ToBeFree (talk) 05:02, 22 March 2020 (UTC)[reply]

Comments at Wikipedia:Help_desk#UnwatchedPages by User:Teratix and User:PrimeHunter reminded me that this page is only open to admins, and from User:Supernerd11's message #Yet another opening proposal above I gather that there have been many proposals to open it wider. Why is that threshold so high? Why don't we - just as an experiment - open it to other editors of good standing, such as rollbackers? That wouldn't require any effort by the developers. To be safe, I would recommend a little effort, though: Before changing the permission, back up the current list. Then, after about two weeks, save the list as it will be current then. Then just compare them and provide a list of all that vanished in the meantime for all to double check. ◄ Sebastian 12:36, 19 February 2020 (UTC)[reply]

I mean it's kind of a hobson's choice. I don't recall exactly what you have to do to be a rollbacker, but it's not that hard I think -- just ask IIRC. And the potential, and temptation, for people to do unseen vandalism is there. I don't see how any small test would prove anything about potential long-term mischief. It's not a great situation, because it'd be nice to get more eyes on these pages, and admins aren't actually better at or more eager to watch pages than other people. I bet there's some solution out there, I just can't think of it. Herostratus (talk) 16:50, 19 February 2020 (UTC)[reply]

What do you mean by “I don't see how any small test would prove anything about potential long-term mischief”. Of course, nobody claims to have any proof, but who needs proof here? There are many ways in which people without admin privileges can damage Wikipedia, but, as I pointed out, this is one that is particularly easy to keep in check. If it turns out after a couple weeks that fears did not materialize and the community decides to make the experiment more permanent, but you still have fears, then nothing prevents you from (even secretly, if you wish!) extending the experiment until you're satisfied. ◄ Sebastian 19:48, 20 February 2020 (UTC)[reply]

I wouldn't mind opening the list to rollbackers. The list entries need watchers who know how to deal with vandalism, and where to report the few rollbackers (or sockpuppets) who misuse their access. The cost of misusing the list is much higher than the cost of watching it. As long as there is a strong inequality between the amount of work needed to harm, and the amount of work needed to protect the encyclopedia, the project will resist attacks. ~ ToBeFree (talk) 01:53, 21 February 2020 (UTC)[reply]

It seems we have consensus here. (Or, Herostratus, do you want to reply above?) Does anyone know how to open this page to others, or do I need to ask at the help desk? ◅ Sebastian 10:12, 3 December 2020 (UTC)[reply]

No, I'm good, it's probably worth a try at least. Herostratus (talk) 13:18, 3 December 2020 (UTC)[reply]

Thank you. I now learned (on User talk:ToBeFree#UnwatchedPages) that “we need a much clearer consensus at a central page like WP:VPPR, or at least via a Request for comments”. I won't get around to that in the next few days, but if no one else starts this, I can do so next week. ◅ Sebastian 14:03, 5 December 2020 (UTC)[reply]

Did anything ever come of this, User:SebastianHelm? Herostratus (talk) 16:57, 1 April 2021 (UTC)m[reply]

From # 731 till the end of the list at # 5000 (which is nowhere near “Z”), I see only talk pages. How can this be? How can a page be watched, but not its talk page? ◅ Sebastian 20:04, 1 December 2020 (UTC)[reply]

Most of them are redirects' talk pages, as the redirect itself doesn't appear in the list. The talk page isn't a redirect, so it appears there. Also, some of them are duplicates: The article and the talk page are unwached, and both appear on the list. The rest are talk page archives (no corresponding article exists), "Good Article" discussions which happen on talk subpages, {{G8-exempt}} talk pages et cetera. ~ ToBeFree (talk) 22:46, 30 December 2020 (UTC)[reply]

Makes sense, thanks! Yes, there are a lot of subpages. Another rather big group are IP address talk pages. (As if we didn't have enough trouble with IP users already! ;-) ) ◅ Sebastian 14:54, 31 December 2020 (UTC)[reply]

Based on the "MediaWiki JS" example at https://www.mediawiki.org/w/index.php?title=API:Watch&oldid=4112591#Sample_code , one may be interested in running the following commands one at a time in their browser's JavaScript console at their own risk.

var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '0',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '500',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '1000',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '1500',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '2000',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '2500',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '3000',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '3500',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '4000',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } );  var params = { action: 'watch', expiry: 'never', generator: 'querypage', gqppage: 'Unwatchedpages', gqplimit: '500', gqpoffset: '4500',  format: 'json'  },  api = new mw.Api(); api.postWithToken( 'watch', params ).done( function ( data ) {  console.log( data ); } ); 

Modify the parameters to fit your needs; documentation is at mw:API:Watch and mw:API:Querypage. There seems to be no way to filter for a specific namespace, so if you want to watchlist articles only, use the "gqplimit" to stop before hitting other namespaces. ~ ToBeFree (talk) 23:28, 30 December 2020 (UTC)[reply]

So, I presume the gqplimit works like the limit parameter in the watchlist? Where is that documented? (I see it in neither of the two MediaWiki pages you linked.)Is there a way to figure out the point when the list hits other namespaces in advance, or does one have to get that via trial and error for each of the commands? ◅ Sebastian 14:55, 31 December 2020 (UTC)[reply]

It's "qplimit" in the Querypage documentation; the "g" is prepended because the module is used as a "g"enerator. It took me a while to get this. The relevant documentation is "Note: Generator parameter names must be prefixed with a 'g', see examples." at mw:API:Watch. Regarding the namespaces, let's say the first "Talk" entry at https://en.wikipedia.org/w/index.php?title=Special:UnwatchedPages&limit=5000 is #740. In this case, you can run the first command unmodified (first 500) and then the second command with a limit of 240 for the rest. The special page's contents are cached for three days, so running a command does not modify the output of the next one. ~ ToBeFree (talk) 02:48, 1 January 2021 (UTC)[reply]

Additional note: As a sysop, you have the apihighlimits user right (Special:ListGroupRights#sysop) and can add up to 5000 in a single command... theoretically. Practically, that just times out and probably creates a noticeable load spike on the database backend. As all other users have a limit of 500 per command, I didn't want to provide copy-and-paste code for using sysop-only limit exceedence. ~ ToBeFree (talk) 02:56, 1 January 2021 (UTC)[reply]