So, I've got some ideas on how to run the drive, especially using existing bots (Enterprisey hasn't replied) to do most of the heavy lifting—now detailed at
Wikipedia:New pages patrol/Backlog drives/October–November 2021. Relying so heavily on the leaderboard for scoring may not work because there may be more than 100 people participating in the drive and/or reviewing articles that month. Let me know what you think about the proposed rules.
Once we get that figured out, I expect the next major thing will be send out a message urging all patrollers to participate in the drive, perhaps about a week before it starts. Then afterward, the coordinators have to add together all the points and award barnstars. (
t ·
c) buidhe08:31, 21 September 2021 (UTC)reply
@
Buidhe: Thanks for setting the initial page up. I like most of the rules, but there are a few things we should probably take note of.
First, the database report includes patrolling redirects. While this is an important thing to do - it's a very different task from regular NPP, and can be done much faster. Using a separate report to only count non-redirects would probably be a good idea.
I like the idea behind re-reviewing - but I'm a bit concerned people might focus heavily on the people at the top of the leaderboard, and end up re-reviewing the same articles. Not sure exactly the solution there, but maybe we could ask people to make sure the article they're re-reviewing hasn't already been re-reviewed by someone else?
Now we have a third coordinator,
Tol has kindly agreed to help out. I agree that the drive should be focusing on articles, not redirects, but is there a database report that excludes redirects? If not, we may have to get people to list the patrolled articles (perhaps on a dedicated subpage for each participant as was done during the AfC drive).
I've added a note to the instructions to avoid duplication with re-reviews.
Database reports are done by the Community Tech team. As far as I can tell, they don't have a reputation for being very responsive. This particular report relies on another bit of code by MusikAnimal. I'm guessing it wouldn't be too complicated to exclude redirects but in order to have two separate reports you would need to get the community tech team to do something. Probably the most likely way to accomplish it is if the NPP community wanted to exclude redirects from the leaderboard calculation either temporarily or permanently, then we could contact MusikAnimal (they are still active). But I don't know if that's desirable. Otherwise, I think we would have to use individual user subpages like the AfC drive did, except we would probably not have a bot to automatically generate them. (
t ·
c) buidhe23:19, 21 September 2021 (UTC)reply
Re: the Community Tech team ... don't have a reputation for being very responsive – I'm assuming you're referring to on-wiki talk pages? There are many dozens, so we do miss things, but we never ignore inquiries when seen!
Phabricator tasks tagged with
#community-tech are certain to be noticed, or you can just ping someone on the team. In this case, you pinged the right person :)
It's fairly simple to change the bot's query to exclude counting pages that are redirects at the time query is ran. However, there's no easy way to query the redirect status at the time the review actually took place (the same reason we gave up on
phab:T157048). So there's a chance a patroller could mark an article as reviewed, then before the next bot run it's changed to a redirect, and hence isn't counted in the report. As long as you're okay with this, I'm happy to make the change. We can also do this on a different page if you want, as a supplement to the current report. — MusikAnimaltalk01:46, 22 September 2021 (UTC)reply
@
MusikAnimal: a separate page would be best - even better would be having a merged report, one for redirects only, and one for non-redirects only (I know redirects-only can be derived from the other two, but that takes more effort than a bot would). And while not perfect that would probably be good enough.
A separate report for merged, redirects-only and non-redirects sounds like a fine idea. And yes, I believe you are correct that patrolled redirects that become deleted are reported as non-redirects, since we can't tell if deleted content is a redirect.
Tol,
Elli I'm really sorry for dropping the ball on this. I don't think it's possible to run the drive as planned. Do you think it would work to put it off until 1 November, running through the end of the month? Are there any updates on the leaderboard issue? (
t ·
c) buidhe02:14, 11 October 2021 (UTC)reply
That sounds fine. I can work on a bot-updated leaderboard. Should it give points for patrols and deletion tagging (via Page Curation)? I know I can monitor those, but draftification would be a bit harder (though I think I can do it).
Tol (
talk |
contribs) @
03:48, 11 October 2021 (UTC)reply
The thing is that all the actions need to be somehow review-able - otherwise someone could game the system by say, draftifying every page they come across. If you could list out every action a person takes - in addition to collecting them all for a leaderboard - that would be great. If that makes sense.
Elli (
talk |
contribs)
13:26, 11 October 2021 (UTC)reply
I think that lists similar to those used in an AfC drive can be a good way to improve accountability and make sure that the system isn't being gamed. However, I could be confused but I thought that was what Tol was proposing to do. The next steps for me to work on are 1) clean up the project page for prime time and 2) draft a mass message, which I will ask a mass message sender to put out to all NPP rights holders sometime during the last week of October. (
t ·
c) buidhe15:52, 11 October 2021 (UTC)reply
Just for information, I have found Page Curation tool to be unreliable in tagging AfD's (it doesn't always complete all the tasks) and so would be inclined to use Twinkle for that to save the bother of having to re-create the deletion discussion manually, page curation is fine for PROD and most CSD's (CSD G12 it only allows one url while twinkle allows up to three) - regular reviewers might be inclined to stick with that
JW 1961Talk22:28, 12 October 2021 (UTC)reply
Hey all. @
Buidhe, sorry I didn't respond on my talk page. It would be a decent amount of work to adapt my code to NPP, and I can't guarantee I'll be available to do it (the drive took a ton of time). I would be happy to run code on my account that needs to see deleted pages (if we ever need to see past deletion tagging) or to help out otherwise (including with bot code, bot approvals, etc), though. I would recommend each re-review page gets its own subpage for reasons of convenience - I didn't see for certain whether that was being done.
Enterprisey (
talk!)
08:29, 17 October 2021 (UTC)reply
Updates?
Hi
Tol, I was just going to update the main page with new scoring based on the tool you were working on, but I wasn't sure whether you had finished it.
In the meantime I've developed a proposal for the message to send out:
@
Buidhe: It's functional, but not perfect. (Also, I intend for it to only edit within its user space, so that it doesn't have to go through a BRFA.) There are still a few problems (mainly to do with deleted articles and redirects). It would just check if the page is an article, a redirect, or has been deleted, and assume that deleted pages were articles (not redirects). I don't have access to deleted contribs, so there's no good way of checking whether a deleted page was a redirect. I'll set it up for a test run (and do some optimisation) around Thursday, and it should be ready then. The mass message looks great!
Tol (
talk |
contribs) @
00:11, 25 October 2021 (UTC)reply
I think giving (maybe more) points for successful deletions would help minimise bias. One could add that to self-reported bonus part, I guess. Usedtobecool☎️02:40, 25 October 2021 (UTC)reply
CSD
quality and depth of patrolling and the use of correct CSD criteria are essential to good reviewing. Will CSD count in this backlog drive? Those articles don't get "marked as reviewed", and most also end up deleted, so it may be difficult to detect. I bring this up not to suggest that it should count (may add too much complexity to whatever bot is doing the tallying), but rather to suggest that maybe we should alter the quoted wording. –
Novem Linguae (
talk)
14:40, 25 October 2021 (UTC)reply
(
edit conflict) @
Novem Linguae: the intent is for CSDs to be counted, yes. I believe this is only done via the bot if the tags are added via the curation sidebar - this isn't ideal, of course. You can see the discussion of this above - not sure if other methods will end up counting - I would think they probably will?
Elli (
talk |
contribs)
14:59, 25 October 2021 (UTC)reply
@
Elli and @
Novem Linguae: Yes, the bot will only record CSD tags using PageCuration. You can also log CSD tags (I recommend Twinkle, as you can turn on a CSD log) and then present it for additional points at the end. I don't believe that any other coordinator is an admin (to view deleted contributions), so it'll be on a basis of trust.
Tol (
talk |
contribs) @
04:14, 26 October 2021 (UTC)reply
The feeling that you've contributed to a collective effort in clearing the backlog. I'm going to participate, and I don't care about getting barnstars. But if they serve as an incentive or even just an amusement, then that's good.
MarioGom (
talk)
17:16, 25 October 2021 (UTC)reply
I encourage you to participate. We started the barnstars at 5 articles because even reviewing a relatively small number helps clear the backlog. (
t ·
c) buidhe21:49, 26 October 2021 (UTC)reply
The subject is the author of a well-reviewed book named The Nay Science: A History of German Indology. Most of the sources cited in the article deal with this single book. Should I request to creator to create book's article and move the relevant content over there ? Or Should I mark this page as reviewed? --
Gazal world (
talk)
19:15, 25 October 2021 (UTC)reply
@
Gazal world: Looking at the article, it doesn't seem like all the sources focus on a single piece of their work - so I'd probably mark it as reviewed (Hans Harder, Angelika Malinar and Thomas Oberlies, in a 2011 editorial for Zeitschrift für Indologie und Südasienstudien on combating "discrimination, racism and sexism", noted that Adluri's works engaged in polemics against multiple German scholars under the veneer of probing ideological orientations of scholarship.[3] wouldn't be that relevant in an article on the 2016 book). If it did, though, I'd ask them to move the page to the book's title and rework it - it's what was done with
Abigail Shrier and tends to work pretty well for authors notable for a single work.
Elli (
talk |
contribs)
21:31, 26 October 2021 (UTC)reply
See, Charm & Strange as well as Vanishing Girls and Bloody Seoul. The reception section contains only these type of sentences: "The book received a starred review from X, positive reviews from Y, and a mediocre review from Z". Is this valid? "Reception" section should have some opinion/remarks/comment by the reviewer. Isn't it? --
Gazal world (
talk)
15:23, 27 October 2021 (UTC)reply
I think that's fine to approve. We don't need the article to be perfect, with that many reviews it will probably be considered notable and it can always be expanded later. (
t ·
c) buidhe21:59, 27 October 2021 (UTC)reply
@
Buidhe and @
Elli: I just noticed four users who have signed up on the
participants list but are not new page patrollers. I'm not sure how to handle this. I think that non-patrollers can assist with patrolling (by tagging pages for deletion, improving new pages, adding cleanup tags, et cetera), but I worry that if they participate in a drive where they get points for tagging pages for deletion and draftifying, this will incentivise them to
err on the side of deletion tagging and draftifying (as leaving alone a good page does not gain any points, whereas a patroller would be able to patrol it and gain a point). What do you think?
Tol (
talk |
contribs) @
20:02, 31 October 2021 (UTC)reply
I think that while they're here in good faith, they should not be allowed to participate without the relevant permission. No offense to them, of course, but the permission is kinda necessary for this.
Elli (
talk |
contribs)
20:05, 31 October 2021 (UTC)reply
I tried using the page curation toolbar for some deletions tonight. I am having some trouble. G5 places the tag then hangs. PROD places the prod, but doesn't log it in my prod log. Perhaps TolBot could be programmed to detect Twinkle in addition to Page Curation. Helpfully, Twinkle edits are
tagged as such. If not, no worries. Thanks. –
Novem Linguae (
talk)
07:50, 1 November 2021 (UTC)reply
I attempted to take a page to AfD using PageCuration, which successfully tagged the page, created an entry in the daily log, notified the creator, but did not create the nomination page itself... Can we allow AfD nominations to be accepted in good faith (like CSD nominations) if we have the permalink of creating the nomination?
Sdrqaz (
talk)
13:01, 1 November 2021 (UTC)reply
NPP flowchart says to mark pages taken to AFD as reviewed, so we may not need a special process for this one. I imagine AFDd pages would count as normal page curations. Correct me if I'm wrong! –
Novem Linguae (
talk)
20:36, 1 November 2021 (UTC)reply
@
Novem Linguae and @
Sdrqaz: Yes, you should patrol pages that you send to AfD. PROD and CSD should be simpler (fewer pages to edit), but Twinkle CSD and PROD logs are acceptable and will be added to your total at the end of the drive. The difference with PageCuration deletion nominations is that they are
logged, whereas Twinkle nominations are not (and, if the nomination is successful, there is no non-deleted record of them unless you turn on the CSD/PROD logs).
Tol (
talk |
contribs) @
21:59, 1 November 2021 (UTC)reply
While we're on the subject of re-reviews, I noticed that
Jessamyn has re-reviewed several redirects. Based on the mass message I received on my talk page, I thought that redirects didn't count. Although the main drive page speaks multiple times of "articles", it might be worth making clear. As for the question of a subpage, please do (and possibly transclude it to the main page). Thanks,
Sdrqaz (
talk)
19:25, 2 November 2021 (UTC)reply
Yeah, while redirects are part of the same workflow, technically, that backlog isn't much of an issue, requires a different amount/type of effort, and is handled by different people than most NPP work.
Elli (
talk |
contribs)
19:55, 2 November 2021 (UTC)reply
Agreed; I think re-reviews should be of articles that were reviewed as part of this drive. I could have my bot make a list of such reviews, if that would be helpful.
Tol (
talk |
contribs) @
21:45, 2 November 2021 (UTC)reply
Do we want to encourage people to use Y and N when re-reviewing, to quickly see which ones pass and fail? Or is the idea to just provide gentle feedback and no need to track things so precisely? –
Novem Linguae (
talk)
21:49, 2 November 2021 (UTC)reply
I'll be honest, I find the re-review workflow really confusing (as someone new to this) while the rest of this process basically makes sense to me. Like, I know how to do a re-review and I definitely understand why it's important, but how to find a page to re-review that hasn't been re-reviewed already seems super challenging. The redirects were just the first pages I found and I didn't know they were redirects until I clicked through. If there was a backlog-November-drive list to work from, that would be excellent.
Jessamyn (
talk)
22:44, 2 November 2021 (UTC)reply
I would personally prefer not to use the Y or N (they might make it look like it was some sort of exam) - the gentle feedback seems sufficient to me
JW 1961Talk23:34, 2 November 2021 (UTC)reply
I think a good idea may be to have the bot generate a list of recent reviews each time it runs, and move reviews there (reply to each item in the list to re-review it). Does this sound good?
Tol (
talk |
contribs) @
00:16, 3 November 2021 (UTC)reply
I'm still working on this, and don't know how long it'll take, as it requires either a structural change to the rest of the code or for me to implement a lot of new stuff. Sorry.
Tol (
talk |
contribs) @
20:59, 6 November 2021 (UTC)reply
@
AntanO: Ah, I see the problem: only the PageCuration toolbar logs things in the page curation (pagetriage-curation) log. I'll change it to use the patrol (patrol) log; that should fix the problem. Just to make sure; does
your patrol log look right?
Tol (
talk |
contribs) @
02:35, 4 November 2021 (UTC)reply
Please be careful using # of patrols as opposed to # of page curations. My experience is that the two rarely correspond 1:1. And one is not completely inclusive of the other. Overall, I believe page curation is the accurate one, and patrol is a random mishmash of people hitting the "mark this page as patrolled" link that appears at the bottom of some namespaces, and certain automatic criteria for patrolling that is executed when performing other actions. I have not been able to figure out what the criteria are. –
Novem Linguae (
talk)
06:31, 4 November 2021 (UTC)reply
I've investigated this, and the patrol log is going to be difficult to use. I'm not sure if I can switch over, and am still investigating exactly how "patrol" works, and looking into edge cases.
Tol (
talk |
contribs) @
02:06, 5 November 2021 (UTC)reply
@
AntanO: Not done: I don't think the "patrol" log can be used, because it contains the title of the page at its first revision, making it impossible to differentiate between patrolling a draft (which can only be done with basic patrol) and patrolling an article which used to be a draft (which can be done either with basic patrol or with PageCuration). "Basic patrol" is the "[Mark this page as patrolled]" button at the bottom. Therefore, I'd like to ask you to please only use PageCuration to patrol pages, so that it can be logged. I can manually add the pages you've already patrolled at the end of the drive.
Tol (
talk |
contribs) @
17:35, 9 November 2021 (UTC)reply
The leaderboard also shows me as '0' reviews, which isn't the case... though I haven't been using the page curation toolbar to do the reviews.
Dan arndt (
talk)
03:40, 4 November 2021 (UTC)reply
I use Firefox and Page Curation and it works fine (I review a small number of redirects and, as it's supposed to be, they don't show up)
JW 1961Talk08:49, 5 November 2021 (UTC)reply
Elli I have been doing that on my more recent reviews but the "leaderboard" still shows me as zero... whereas I would have done at least one hundred or more... :-(
Dan arndt (
talk)
04:25, 9 November 2021 (UTC)reply
@
Dan arndt: still doesn't show them in your log. Can you give me an example of a page you marked as reviewed in this way so I could check its current status and the relevant logs?
Elli (
talk |
contribs)
05:34, 9 November 2021 (UTC)reply
@
Dan arndt: ah, you're doing articles for creation, not new page patrol. New page patrol is for pages that are already in the article namespace. At
Special:NewPagesFeed, make sure "New Page Patrol" is selected in the top-left of the grey box, not "Articles for Creation". Then you should see articles and be able to review them.
Elli (
talk |
contribs)
14:48, 9 November 2021 (UTC)reply
Sorry, I didn't see the ping earlier, yes you made the right call. I have no difficulties with any of my reviews ever being undone by another reviewer if I miss out on something. Thank you
JW 1961Talk18:04, 7 November 2021 (UTC)reply
Hi @
Tol: it looks like TolBot has not made any substantial updates to the leaderboard since 11 hours ago, even though it's the last day of the backlog drive. This morning it decreased a single participant's number from 17 to 16, and then self-reverted two hours later. Any ideas? Thanks for all the technical work you put into it!
DanCherek (
talk)
14:09, 30 November 2021 (UTC)reply
Barnstars will be awarded based on
this chart in my sandbox. If there are any articles that you reviewed that aren't reflected in this chart, please set me know sooner rather than later. My plan is to send out barnstars tomorrow. (
t ·
c) buidhe06:24, 1 December 2021 (UTC)reply
Buidhe, I don't think your sandbox takes into account "We will also accept CSD logs made by Twinkle, which will be added to the leaderboard at the end of the drive, based on the principle of good faith."Sdrqaz (
talk)
08:18, 1 December 2021 (UTC)reply
Except for a few participants who have contacted me, that list does not include CSD logs at present. That's why I'm inviting participants to submit any relevant logs that are missing from their totals. (
t ·
c) buidhe08:31, 1 December 2021 (UTC)reply
There's some interest in a new backlog drive.
Tol, how difficult would it be to run your script again, say we were doing a drive in mid-June or July? I can work on manually setting up pages as necessary and sending out messages. (
t ·
c) buidhe19:28, 22 May 2022 (UTC)reply
@
Buidhe, the script I wrote was cobbled together somewhat quickly, but it should easily run again without too much work. I should probably take this opportunity to make it work better and more elegantly, though. I'd be available to handle the script/bot side of this again!
Tol (
talk |
contribs) @
19:47, 22 May 2022 (UTC)reply
So, I've got some ideas on how to run the drive, especially using existing bots (Enterprisey hasn't replied) to do most of the heavy lifting—now detailed at
Wikipedia:New pages patrol/Backlog drives/October–November 2021. Relying so heavily on the leaderboard for scoring may not work because there may be more than 100 people participating in the drive and/or reviewing articles that month. Let me know what you think about the proposed rules.
Once we get that figured out, I expect the next major thing will be send out a message urging all patrollers to participate in the drive, perhaps about a week before it starts. Then afterward, the coordinators have to add together all the points and award barnstars. (
t ·
c) buidhe08:31, 21 September 2021 (UTC)reply
@
Buidhe: Thanks for setting the initial page up. I like most of the rules, but there are a few things we should probably take note of.
First, the database report includes patrolling redirects. While this is an important thing to do - it's a very different task from regular NPP, and can be done much faster. Using a separate report to only count non-redirects would probably be a good idea.
I like the idea behind re-reviewing - but I'm a bit concerned people might focus heavily on the people at the top of the leaderboard, and end up re-reviewing the same articles. Not sure exactly the solution there, but maybe we could ask people to make sure the article they're re-reviewing hasn't already been re-reviewed by someone else?
Now we have a third coordinator,
Tol has kindly agreed to help out. I agree that the drive should be focusing on articles, not redirects, but is there a database report that excludes redirects? If not, we may have to get people to list the patrolled articles (perhaps on a dedicated subpage for each participant as was done during the AfC drive).
I've added a note to the instructions to avoid duplication with re-reviews.
Database reports are done by the Community Tech team. As far as I can tell, they don't have a reputation for being very responsive. This particular report relies on another bit of code by MusikAnimal. I'm guessing it wouldn't be too complicated to exclude redirects but in order to have two separate reports you would need to get the community tech team to do something. Probably the most likely way to accomplish it is if the NPP community wanted to exclude redirects from the leaderboard calculation either temporarily or permanently, then we could contact MusikAnimal (they are still active). But I don't know if that's desirable. Otherwise, I think we would have to use individual user subpages like the AfC drive did, except we would probably not have a bot to automatically generate them. (
t ·
c) buidhe23:19, 21 September 2021 (UTC)reply
Re: the Community Tech team ... don't have a reputation for being very responsive – I'm assuming you're referring to on-wiki talk pages? There are many dozens, so we do miss things, but we never ignore inquiries when seen!
Phabricator tasks tagged with
#community-tech are certain to be noticed, or you can just ping someone on the team. In this case, you pinged the right person :)
It's fairly simple to change the bot's query to exclude counting pages that are redirects at the time query is ran. However, there's no easy way to query the redirect status at the time the review actually took place (the same reason we gave up on
phab:T157048). So there's a chance a patroller could mark an article as reviewed, then before the next bot run it's changed to a redirect, and hence isn't counted in the report. As long as you're okay with this, I'm happy to make the change. We can also do this on a different page if you want, as a supplement to the current report. — MusikAnimaltalk01:46, 22 September 2021 (UTC)reply
@
MusikAnimal: a separate page would be best - even better would be having a merged report, one for redirects only, and one for non-redirects only (I know redirects-only can be derived from the other two, but that takes more effort than a bot would). And while not perfect that would probably be good enough.
A separate report for merged, redirects-only and non-redirects sounds like a fine idea. And yes, I believe you are correct that patrolled redirects that become deleted are reported as non-redirects, since we can't tell if deleted content is a redirect.
Tol,
Elli I'm really sorry for dropping the ball on this. I don't think it's possible to run the drive as planned. Do you think it would work to put it off until 1 November, running through the end of the month? Are there any updates on the leaderboard issue? (
t ·
c) buidhe02:14, 11 October 2021 (UTC)reply
That sounds fine. I can work on a bot-updated leaderboard. Should it give points for patrols and deletion tagging (via Page Curation)? I know I can monitor those, but draftification would be a bit harder (though I think I can do it).
Tol (
talk |
contribs) @
03:48, 11 October 2021 (UTC)reply
The thing is that all the actions need to be somehow review-able - otherwise someone could game the system by say, draftifying every page they come across. If you could list out every action a person takes - in addition to collecting them all for a leaderboard - that would be great. If that makes sense.
Elli (
talk |
contribs)
13:26, 11 October 2021 (UTC)reply
I think that lists similar to those used in an AfC drive can be a good way to improve accountability and make sure that the system isn't being gamed. However, I could be confused but I thought that was what Tol was proposing to do. The next steps for me to work on are 1) clean up the project page for prime time and 2) draft a mass message, which I will ask a mass message sender to put out to all NPP rights holders sometime during the last week of October. (
t ·
c) buidhe15:52, 11 October 2021 (UTC)reply
Just for information, I have found Page Curation tool to be unreliable in tagging AfD's (it doesn't always complete all the tasks) and so would be inclined to use Twinkle for that to save the bother of having to re-create the deletion discussion manually, page curation is fine for PROD and most CSD's (CSD G12 it only allows one url while twinkle allows up to three) - regular reviewers might be inclined to stick with that
JW 1961Talk22:28, 12 October 2021 (UTC)reply
Hey all. @
Buidhe, sorry I didn't respond on my talk page. It would be a decent amount of work to adapt my code to NPP, and I can't guarantee I'll be available to do it (the drive took a ton of time). I would be happy to run code on my account that needs to see deleted pages (if we ever need to see past deletion tagging) or to help out otherwise (including with bot code, bot approvals, etc), though. I would recommend each re-review page gets its own subpage for reasons of convenience - I didn't see for certain whether that was being done.
Enterprisey (
talk!)
08:29, 17 October 2021 (UTC)reply
Updates?
Hi
Tol, I was just going to update the main page with new scoring based on the tool you were working on, but I wasn't sure whether you had finished it.
In the meantime I've developed a proposal for the message to send out:
@
Buidhe: It's functional, but not perfect. (Also, I intend for it to only edit within its user space, so that it doesn't have to go through a BRFA.) There are still a few problems (mainly to do with deleted articles and redirects). It would just check if the page is an article, a redirect, or has been deleted, and assume that deleted pages were articles (not redirects). I don't have access to deleted contribs, so there's no good way of checking whether a deleted page was a redirect. I'll set it up for a test run (and do some optimisation) around Thursday, and it should be ready then. The mass message looks great!
Tol (
talk |
contribs) @
00:11, 25 October 2021 (UTC)reply
I think giving (maybe more) points for successful deletions would help minimise bias. One could add that to self-reported bonus part, I guess. Usedtobecool☎️02:40, 25 October 2021 (UTC)reply
CSD
quality and depth of patrolling and the use of correct CSD criteria are essential to good reviewing. Will CSD count in this backlog drive? Those articles don't get "marked as reviewed", and most also end up deleted, so it may be difficult to detect. I bring this up not to suggest that it should count (may add too much complexity to whatever bot is doing the tallying), but rather to suggest that maybe we should alter the quoted wording. –
Novem Linguae (
talk)
14:40, 25 October 2021 (UTC)reply
(
edit conflict) @
Novem Linguae: the intent is for CSDs to be counted, yes. I believe this is only done via the bot if the tags are added via the curation sidebar - this isn't ideal, of course. You can see the discussion of this above - not sure if other methods will end up counting - I would think they probably will?
Elli (
talk |
contribs)
14:59, 25 October 2021 (UTC)reply
@
Elli and @
Novem Linguae: Yes, the bot will only record CSD tags using PageCuration. You can also log CSD tags (I recommend Twinkle, as you can turn on a CSD log) and then present it for additional points at the end. I don't believe that any other coordinator is an admin (to view deleted contributions), so it'll be on a basis of trust.
Tol (
talk |
contribs) @
04:14, 26 October 2021 (UTC)reply
The feeling that you've contributed to a collective effort in clearing the backlog. I'm going to participate, and I don't care about getting barnstars. But if they serve as an incentive or even just an amusement, then that's good.
MarioGom (
talk)
17:16, 25 October 2021 (UTC)reply
I encourage you to participate. We started the barnstars at 5 articles because even reviewing a relatively small number helps clear the backlog. (
t ·
c) buidhe21:49, 26 October 2021 (UTC)reply
The subject is the author of a well-reviewed book named The Nay Science: A History of German Indology. Most of the sources cited in the article deal with this single book. Should I request to creator to create book's article and move the relevant content over there ? Or Should I mark this page as reviewed? --
Gazal world (
talk)
19:15, 25 October 2021 (UTC)reply
@
Gazal world: Looking at the article, it doesn't seem like all the sources focus on a single piece of their work - so I'd probably mark it as reviewed (Hans Harder, Angelika Malinar and Thomas Oberlies, in a 2011 editorial for Zeitschrift für Indologie und Südasienstudien on combating "discrimination, racism and sexism", noted that Adluri's works engaged in polemics against multiple German scholars under the veneer of probing ideological orientations of scholarship.[3] wouldn't be that relevant in an article on the 2016 book). If it did, though, I'd ask them to move the page to the book's title and rework it - it's what was done with
Abigail Shrier and tends to work pretty well for authors notable for a single work.
Elli (
talk |
contribs)
21:31, 26 October 2021 (UTC)reply
See, Charm & Strange as well as Vanishing Girls and Bloody Seoul. The reception section contains only these type of sentences: "The book received a starred review from X, positive reviews from Y, and a mediocre review from Z". Is this valid? "Reception" section should have some opinion/remarks/comment by the reviewer. Isn't it? --
Gazal world (
talk)
15:23, 27 October 2021 (UTC)reply
I think that's fine to approve. We don't need the article to be perfect, with that many reviews it will probably be considered notable and it can always be expanded later. (
t ·
c) buidhe21:59, 27 October 2021 (UTC)reply
@
Buidhe and @
Elli: I just noticed four users who have signed up on the
participants list but are not new page patrollers. I'm not sure how to handle this. I think that non-patrollers can assist with patrolling (by tagging pages for deletion, improving new pages, adding cleanup tags, et cetera), but I worry that if they participate in a drive where they get points for tagging pages for deletion and draftifying, this will incentivise them to
err on the side of deletion tagging and draftifying (as leaving alone a good page does not gain any points, whereas a patroller would be able to patrol it and gain a point). What do you think?
Tol (
talk |
contribs) @
20:02, 31 October 2021 (UTC)reply
I think that while they're here in good faith, they should not be allowed to participate without the relevant permission. No offense to them, of course, but the permission is kinda necessary for this.
Elli (
talk |
contribs)
20:05, 31 October 2021 (UTC)reply
I tried using the page curation toolbar for some deletions tonight. I am having some trouble. G5 places the tag then hangs. PROD places the prod, but doesn't log it in my prod log. Perhaps TolBot could be programmed to detect Twinkle in addition to Page Curation. Helpfully, Twinkle edits are
tagged as such. If not, no worries. Thanks. –
Novem Linguae (
talk)
07:50, 1 November 2021 (UTC)reply
I attempted to take a page to AfD using PageCuration, which successfully tagged the page, created an entry in the daily log, notified the creator, but did not create the nomination page itself... Can we allow AfD nominations to be accepted in good faith (like CSD nominations) if we have the permalink of creating the nomination?
Sdrqaz (
talk)
13:01, 1 November 2021 (UTC)reply
NPP flowchart says to mark pages taken to AFD as reviewed, so we may not need a special process for this one. I imagine AFDd pages would count as normal page curations. Correct me if I'm wrong! –
Novem Linguae (
talk)
20:36, 1 November 2021 (UTC)reply
@
Novem Linguae and @
Sdrqaz: Yes, you should patrol pages that you send to AfD. PROD and CSD should be simpler (fewer pages to edit), but Twinkle CSD and PROD logs are acceptable and will be added to your total at the end of the drive. The difference with PageCuration deletion nominations is that they are
logged, whereas Twinkle nominations are not (and, if the nomination is successful, there is no non-deleted record of them unless you turn on the CSD/PROD logs).
Tol (
talk |
contribs) @
21:59, 1 November 2021 (UTC)reply
While we're on the subject of re-reviews, I noticed that
Jessamyn has re-reviewed several redirects. Based on the mass message I received on my talk page, I thought that redirects didn't count. Although the main drive page speaks multiple times of "articles", it might be worth making clear. As for the question of a subpage, please do (and possibly transclude it to the main page). Thanks,
Sdrqaz (
talk)
19:25, 2 November 2021 (UTC)reply
Yeah, while redirects are part of the same workflow, technically, that backlog isn't much of an issue, requires a different amount/type of effort, and is handled by different people than most NPP work.
Elli (
talk |
contribs)
19:55, 2 November 2021 (UTC)reply
Agreed; I think re-reviews should be of articles that were reviewed as part of this drive. I could have my bot make a list of such reviews, if that would be helpful.
Tol (
talk |
contribs) @
21:45, 2 November 2021 (UTC)reply
Do we want to encourage people to use Y and N when re-reviewing, to quickly see which ones pass and fail? Or is the idea to just provide gentle feedback and no need to track things so precisely? –
Novem Linguae (
talk)
21:49, 2 November 2021 (UTC)reply
I'll be honest, I find the re-review workflow really confusing (as someone new to this) while the rest of this process basically makes sense to me. Like, I know how to do a re-review and I definitely understand why it's important, but how to find a page to re-review that hasn't been re-reviewed already seems super challenging. The redirects were just the first pages I found and I didn't know they were redirects until I clicked through. If there was a backlog-November-drive list to work from, that would be excellent.
Jessamyn (
talk)
22:44, 2 November 2021 (UTC)reply
I would personally prefer not to use the Y or N (they might make it look like it was some sort of exam) - the gentle feedback seems sufficient to me
JW 1961Talk23:34, 2 November 2021 (UTC)reply
I think a good idea may be to have the bot generate a list of recent reviews each time it runs, and move reviews there (reply to each item in the list to re-review it). Does this sound good?
Tol (
talk |
contribs) @
00:16, 3 November 2021 (UTC)reply
I'm still working on this, and don't know how long it'll take, as it requires either a structural change to the rest of the code or for me to implement a lot of new stuff. Sorry.
Tol (
talk |
contribs) @
20:59, 6 November 2021 (UTC)reply
@
AntanO: Ah, I see the problem: only the PageCuration toolbar logs things in the page curation (pagetriage-curation) log. I'll change it to use the patrol (patrol) log; that should fix the problem. Just to make sure; does
your patrol log look right?
Tol (
talk |
contribs) @
02:35, 4 November 2021 (UTC)reply
Please be careful using # of patrols as opposed to # of page curations. My experience is that the two rarely correspond 1:1. And one is not completely inclusive of the other. Overall, I believe page curation is the accurate one, and patrol is a random mishmash of people hitting the "mark this page as patrolled" link that appears at the bottom of some namespaces, and certain automatic criteria for patrolling that is executed when performing other actions. I have not been able to figure out what the criteria are. –
Novem Linguae (
talk)
06:31, 4 November 2021 (UTC)reply
I've investigated this, and the patrol log is going to be difficult to use. I'm not sure if I can switch over, and am still investigating exactly how "patrol" works, and looking into edge cases.
Tol (
talk |
contribs) @
02:06, 5 November 2021 (UTC)reply
@
AntanO: Not done: I don't think the "patrol" log can be used, because it contains the title of the page at its first revision, making it impossible to differentiate between patrolling a draft (which can only be done with basic patrol) and patrolling an article which used to be a draft (which can be done either with basic patrol or with PageCuration). "Basic patrol" is the "[Mark this page as patrolled]" button at the bottom. Therefore, I'd like to ask you to please only use PageCuration to patrol pages, so that it can be logged. I can manually add the pages you've already patrolled at the end of the drive.
Tol (
talk |
contribs) @
17:35, 9 November 2021 (UTC)reply
The leaderboard also shows me as '0' reviews, which isn't the case... though I haven't been using the page curation toolbar to do the reviews.
Dan arndt (
talk)
03:40, 4 November 2021 (UTC)reply
I use Firefox and Page Curation and it works fine (I review a small number of redirects and, as it's supposed to be, they don't show up)
JW 1961Talk08:49, 5 November 2021 (UTC)reply
Elli I have been doing that on my more recent reviews but the "leaderboard" still shows me as zero... whereas I would have done at least one hundred or more... :-(
Dan arndt (
talk)
04:25, 9 November 2021 (UTC)reply
@
Dan arndt: still doesn't show them in your log. Can you give me an example of a page you marked as reviewed in this way so I could check its current status and the relevant logs?
Elli (
talk |
contribs)
05:34, 9 November 2021 (UTC)reply
@
Dan arndt: ah, you're doing articles for creation, not new page patrol. New page patrol is for pages that are already in the article namespace. At
Special:NewPagesFeed, make sure "New Page Patrol" is selected in the top-left of the grey box, not "Articles for Creation". Then you should see articles and be able to review them.
Elli (
talk |
contribs)
14:48, 9 November 2021 (UTC)reply
Sorry, I didn't see the ping earlier, yes you made the right call. I have no difficulties with any of my reviews ever being undone by another reviewer if I miss out on something. Thank you
JW 1961Talk18:04, 7 November 2021 (UTC)reply
Hi @
Tol: it looks like TolBot has not made any substantial updates to the leaderboard since 11 hours ago, even though it's the last day of the backlog drive. This morning it decreased a single participant's number from 17 to 16, and then self-reverted two hours later. Any ideas? Thanks for all the technical work you put into it!
DanCherek (
talk)
14:09, 30 November 2021 (UTC)reply
Barnstars will be awarded based on
this chart in my sandbox. If there are any articles that you reviewed that aren't reflected in this chart, please set me know sooner rather than later. My plan is to send out barnstars tomorrow. (
t ·
c) buidhe06:24, 1 December 2021 (UTC)reply
Buidhe, I don't think your sandbox takes into account "We will also accept CSD logs made by Twinkle, which will be added to the leaderboard at the end of the drive, based on the principle of good faith."Sdrqaz (
talk)
08:18, 1 December 2021 (UTC)reply
Except for a few participants who have contacted me, that list does not include CSD logs at present. That's why I'm inviting participants to submit any relevant logs that are missing from their totals. (
t ·
c) buidhe08:31, 1 December 2021 (UTC)reply
There's some interest in a new backlog drive.
Tol, how difficult would it be to run your script again, say we were doing a drive in mid-June or July? I can work on manually setting up pages as necessary and sending out messages. (
t ·
c) buidhe19:28, 22 May 2022 (UTC)reply
@
Buidhe, the script I wrote was cobbled together somewhat quickly, but it should easily run again without too much work. I should probably take this opportunity to make it work better and more elegantly, though. I'd be available to handle the script/bot side of this again!
Tol (
talk |
contribs) @
19:47, 22 May 2022 (UTC)reply