Wikipedia talk:New pages patrol/Unpatrolled articles

Back of the unpatrolled backlog
Thread moved from Wikipedia talk:New pages patrol on 18:15, 26 January 2011 (UTC)

Anyone working the back of the unpatrolled backlog? I'm sure there's a few people chipping away at the coal face there. :) Malcolmxl5 (talk) 18:14, 18 December 2010 (UTC)


 * I have been on and off for the past three months or so, but only to prove a point. I  usually work on  clearing  the whole last  day  in  one session  if it's feasible. Anything  between 50 and 300 a day  get through  without  being  vetted. I feel the work I'm doing on it is a waste of time. I  feel that those that  get  left  untouched are the ones the front-end patrollers leave as too difficult, which  kind of makes a mock  of the system's objectives.--Kudpung (talk) 06:18, 25 December 2010 (UTC)
 * Well, any article getting reviewed is still a good thing, and there are actually editorial opportunities, but once articles get moved out at the back, it gets indeed discouraging. Nevertheless there were recently periods when nothing left without getting patrolled. A more visible marker might help, maybe updated by a bot and including an edit summary, to indicate how close we are the 20 day threshold. --Tikiwont (talk) 14:20, 28 December 2010 (UTC)
 * Don't be discouraged, Kudpung, your work and that of others at the back-end is valued. :-) I usually try and do some everyday (though not over Xmas, I'm afraid) but I had no idea who else is working away there. It would be good to find ways of managing the backlog effectively; a bot sounds like a good idea, Tikiwont, perhaps notifying members of a taskforce of progress? I wonder if there some merit in taking a coordinated approach to chipping away at the back there? Could we organise 10 people to patrol 50 pages at the back-end each day?  Malcolmxl5 (talk) 23:32, 28 December 2010 (UTC)


 * I've already been putting some feelers and suggestions out. The first thing to do is to give them a traceable collar - a bot that puts all the unpatrolled pages over 30 days into an invisible category. This is uncontroversial and needs no discussion, and I already have a bot handler lined up. The bot needs to run every 24 hours at UTC, apply the cat, and send the list to a destination page here at WP:NP. We can then later discuss what happens to them. I think it's unlikely that there will be a great  response to a drive to clean up the backlog - I'd like very much to be proven wrong. Except when I am systematically  cleaning up a whole day's backlog, I tend to cherry pick through the rest of the days, for example, BLP schools, and bands, although I  very rarely touch sportspeople in  the assumption  that people from the footy and American ball  game projects  know more about it than I do. --Kudpung (talk) 02:15, 29 December 2010 (UTC)


 * We're waiting for Kingpin to chime in with some advice on  bot creation. What  is needed is a bot to run every 24 hrs, tag the pages with an invisible cat, and put a daily  report  on a destination page: New pages patrol/30-day list. I'm sure that  it's not going  to be as technically simple as I envisage, because I believe each page gets pushed off at  exactly 72 hours. Kudpung (talk) 03:40, 29 December 2010 (UTC)
 * The only slight flaw is that if you run it every 24 hours on articles unpatrolled after more than 696 hours, you will have up to 24 hours where articles are both marked as unpatrolled and marked with a hidden cat. Of course if the bot marks them as patrolled or if the bot runs continually you could avoid that.  Ϣere Spiel  Chequers  09:22, 29 December 2010 (UTC)
 * It would be possible to get around that problem by collecting the pages to tag a day early, and then tagging them a day late (when they've already fallen off the back of the log) but re-checking that they haven't been manually patrolled in the meantime. That would also deal with the problem of having to worry about getting the pages just in time. One thing I'm wondering at the moment is how you plan to keep this category clean, eventually it's surely going to just end up overloaded with pages which aren't being removed from it (partly because users who are technically reviewing the pages don't remove the category, and partly because there are too many). Do we actually want multiple categories named by date? - Kingpin13 (talk) 13:04, 29 December 2010 (UTC)
 * Sub categories by month and year would help. As for how we keep this clean, I think that is part of a broader issue re backlogs and recruitment and retention of editors. The key thing is that if we can't identify the backlogs we have no way of dealing with it. Once we've identified the issue we have the option of doing something about it.  Ϣere Spiel  Chequers  13:24, 29 December 2010 (UTC)
 * Yes, as I mentioned before, what we do with them once they have been tagged with  an invisible cat is going to be the subject of discussion because it will have a site-wide impact. For the moment,  as many of them are untraceable due to having neither cats, maintenance templates,  nor project templates, they are largely untraceable until stumbled upon by accident. The most urgent thing is to put a  collar on them so we can keep them on the radar. If they are each removed individually from the list at 720 hours, then Kingpin's solution  of a double bot  run  may  be the only  solution, although  somewhat  complicated. Perhaps if the yellow pages were reprogrammed so that  one 24 hour's worth (of one date) were pushed into the void together, then a bot  could catch  them  in  one pass. Kudpung (talk) 09:47, 1 January 2011 (UTC)


 * The backlog seems to have gained a 3-day respite over the Xmas/New year holiday. This could be due to  either a drop in new articles over that period or a greater involvement of patrollers during  their holiday time, or both. I admit I haven't done much myself over the same period, but  when I  have cleared off a day or two's backlog in  the past, it  always crept back. I think it  would be a good idea if we started labelling  them as soon  as possible.Kudpung (talk) 05:32, 7 January 2011 (UTC)
 * Hmmm, I'm still not convinced an invisible category is better than a template at the top of the page (e.g. unreviewed), since users are more likely to remove the template when reviewing the page. I'll try looking at some code when I get the chance. However, the two main problems at this time, is that I'm still not entirely clear on how the end of the NewPages log works. And also, having cleared up a 3-day respite will actually make it more difficult to test the bot. Also, if approval is going to take to long, it would be possible to log the pages to my (or the bot's) user space in the mean time. - Kingpin13 (talk) 06:46, 7 January 2011 (UTC)
 * Article authors are also more likely to remove a unreviewed template from the top of the article they created to prevent it from being nominated for deletion. Is there any way for the bot to ascertain whether the article has truly been marked as patrolled, and if not, to restore the unreviewed template?  Also, wouldn't it be easier and/or more efficient for a toolserver script or new developer tool to create a list of articles that have an empty patrol log, rather than creating a bot that is constantly monitoring the end of the newpages log?  I've got some Python and Javascript experience and have been thinking about trying to get a toolserver account, unless someone else is already working on something like this?  Snotty Wong   comment 14:49, 7 January 2011 (UTC)

The problem is, Snotty, that the articles are there because they are mainly made by SPA or mass-stub maniacs who never return  to  the page to  even see if it's still there. Long term, I  think  we're looking  at  the suggestions  I made on your talk  page, but  those are going to be major site-wide decisions and will take a year or more in  discussion  to  get  agreed and another 6 months  to  get  implemented. In the meantime, what  we need is to  get  the things tagged. First off, it  will give us a base to  make  stats from  to  support our claims for a big solution. I've already made a destination  page here somewhere for the list. Kingpin needs to  find out who  programmed the yellow pages and get  a bot  together - and even that may  take months to  get  approved even if KP  can  design it  quickly. I think a hidden  cat  is the immediate answer - if  we want  a page-top  template, it will  need even more discussion,  design  work, and consensus. Just making  a bot to  feed a destination  page list based on  an invisible cat seems to  be the way  to  go. Clean up  drives could be organised later. I'll let  you  guys worry  about  the technicalities ;)  --Kudpung (talk) 15:06, 7 January 2011 (UTC)
 * I think it may have been User:Gmaxwell who programmed the yellow pages. Malcolmxl5 (talk) 23:16, 7 January 2011 (UTC)
 * Gmax hasn't been editing for 4 months. I've left a message on his tp and sent him an email. --Kudpung (talk) 06:34, 8 January 2011 (UTC)
 * I've been making decent progress on a bot script. I've got a script that can quickly grab all of the new pages for a single day, now I just need to perfect the mechanism for checking whether they're patrolled.  I'll let you know if I come up with anything.  I think a toolserver script would be more efficient, but toolserver is over my head.  Snotty Wong   verbalize 16:24, 8 January 2011 (UTC)
 * Ok, I've got a script that can do the following. It's all somewhat manual at this point, but it will gradually become more automated.  Here's what it can do:
 * Grab a list of all articles created on a certain day, store it in a text file.
 * Once that day has expired at Special:Newpages (i.e. after 30 days), it can go through the list and check each article for patrolled status, and store a list of any unpatrolled articles in another text file.
 * I can then open that text file, and copy and paste the list for a particular day into a page made for that purpose.
 * The script doesn't actually make any edits at this point, and I wouldn't be comfortable with it doing so unless I applied for bot status. If I got bot status, however, it would be trivial for the script to add a Category:Expired unpatrolled articles to the end of each article it finds to be unpatrolled, and then we could have a category page instead of copying and pasting the names from a text file.  Then, whoever patrols the article would simply have to remove that hidden category from the page (and we could ask the keepers of Twinkle to add that functionality to their script).  I'll try applying for bot status and see how that goes.  Somehow, we're about 4 days ahead of the backlog at this point, so I won't be able to fully test the script until the backlog creeps back up.  Snotty Wong   express 17:21, 8 January 2011 (UTC)
 * Applied for bot status. In the meantime, if the backlog creeps back up, I can have the script write the list of unpatrolled articles to the bot's talk page.  It appears that is "legal" even if you don't have bot approval yet.  I'll update the code to make that happen, and I'll test it out next time the backlog is exceeded.  Snotty Wong   squeal 18:22, 8 January 2011 (UTC)

Discussion on the details
I got trial approval of the bot (see here). I'd like to start a discussion to find a consensus on exactly what the best way of tracking these articles will be. We have at least a few days before the backlog catches up with us again. I think we need to discuss the following questions: Any other details we need to iron out? At this point, I have compiled a list of all articles created per day from Dec 13th to Dec 31st. In the event the backlog catches up with us before Jan 31st (pretty likely), I will run the bot to check the articles for patrolled status. I'll wait until at least the day after they have been dropped from Special:Newpages before categorizing them. Snotty Wong  express 18:55, 9 January 2011 (UTC)
 * What category/categories should expired unpatrolled articles be added to?
 * Should it be just one large category, or should there be a separate category for each day, or each week, or each month?
 * The discussion above indicates there is some interest in posting a list of articles to either a subpage of NPP or in the bot's userspace (in addition to categorizing them). Should the bot do this, and if so, where is the best place to post these lists?  Also, why does this need to be done?
 * Once categories have been created and expired unpatrolled articles have been exposed, new page patrollers can try to clear the backlog by going through the articles in the new category. However, when they click the link to go to an article (from the category page), the article won't be shown in "patrol mode", i.e. there won't be a "Mark as patrolled" button in the bottom right.  What is the best way to resolve this issue?
 * First of all, thanks for the effort. In general I like the list thing, mostly because I think that the main objective is to get things patrolled in time. Having more detailed information on how much and what falls out is interesting and may stimulate to avoid it happen in the first place or improve next month; (For example we'd know that say in January 2011 we patrolled 95% of new pages and these 85 would be the ones left out.) If such articles are processed further even better, but moving articles into categories (why not simply into Category:Unreviewed new articles? There could be a filter on creators removing the tag themselves.) - also moves the problem elsewhere. Moreover with categories alone, it is difficult to know later what was inside and why or by whom it was moved out.--Tikiwont (talk) 19:32, 9 January 2011 (UTC)
 * My understanding was that the goal of this process was two-fold: firstly, to gain information on statistics of how many articles fall off the newpages list without being patrolled, and secondly, to provide a way of tracking these articles such that they can be patrolled further out than 30 days from creation.  It seems that compiling a list of articles facilitates the former goal, while putting them in categories facilitates the latter.  If we only keep a list of articles, then we will rely on editors to "cross articles off the list" when they have patrolled them, which could potentially affect our statistics count.  With the category, the editor needs to merely delete the category from the page when patrolling, and we could even make it a very visible template, like REMOVE THIS TEMPLATE WHEN PATROLLING THIS ARTICLE (and all the template would do would add the article to the hidden category).  However, the problem remains that clicking on a link in a category page will never bring you to the article in such a way that you have the "Mark as patrolled" link in the bottom right.  I've been messing around and I believe I've found a way to create a link that will bring you to the page with the patrol link.  So, I suppose I could have the bot create lists of articles with those types of links, and then perhaps run the bot again each day to check if any pages which are in the category have been patrolled, and then automatically remove them from the category.  But this is getting quite cumbersome.  Surely there is a better way to go about this... Snotty Wong   spill the beans 19:44, 9 January 2011 (UTC)
 * Well I wanted to emphasize the overlying objective of getting things patrolled in the first place where I personally see numbers and list instrumental; If things are not patrolled it is also in my opinion good to follow up, but I just want avoid that we try to mimic the same thing belatedly and in a more cumbersome way. Rather I see it then as something different than patrolling, in which case I'd simply go along with something existing and simple such as adding the unreviewed tag which has a monthly basis already,--Tikiwont (talk) 20:16, 9 January 2011 (UTC)

{ec}You've done some excellent work already Snotty! OK, so one-by-one: Kudpung (talk) 19:58, 9 January 2011 (UTC)
 * Cat: I would go  along  with  the suggestion  of Category:Unreviewed new articles
 * Assuming there could be up  to  300 pages per fay (there have been in  the past), perhaps a daily  category;
 * The subpage to post  to  is:  New pages patrol/30-day list. This is needed becuse that  is where anyone will go  who  wants to  work  on the backlog. they  won't  want  to  sift  through  cats. It  will  look  something  like the AlertBot  reports. People and backlog projects could transcude the list  to  their own pages.
 * There won't be the 'unpatrolled' tag on  them, but  the bot  will have left  a discrete mark on  them  that  can  be removed when the problem  page has been addressed by  an editor. Whether we make it invisible is another decision.

Yes, the goal is kind of two-fold. The first objective is to somehow put a magnetic collar on the dogs that will be stray after 30 days, let loose to prowl without apedigree, and the problem is that many of these pages neither have cats, stubs, maintenance templates, nor project templates on the talk pages, thus nothing they can be traced with.

The second objective preempts the lengthy discussion about needing to prove why more drastic measure may need to be taken on what to do with the ones that in spite of all efforts to get them to comply with our notability and sourcing policies, fail to meet those requirements. I've outlined some of those possible solutions on you talk  page. The statistics would be an important piece of evidence. --Kudpung (talk) 20:37, 9 January 2011 (UTC)


 * Ok, so just to clarify, the bot will do two things when it finds an unpatrolled article that has fallen off the Special:Newpages list:
 * Add New unreviewed article to the top of the article, with the appropriate date.
 * Post a link to the article at New pages patrol/30-day list, preferably a link that would bring you to the article with a "Mark this page as patrolled" link, if possible.
 * Also, the way the bot is written, it will automatically keep records (on my hard drive) of how many articles fell off the backlog each day, so I should be able to easily compile statistics from that data when the time comes (rather than relying on the 30-day list, which I'm still a bit unsure exactly how it will be used). Snotty Wong   prattle 23:38, 9 January 2011 (UTC)
 * I think I got the link problem figured out. See User:Snotbot/sandbox.  The links on that page will bring you to the article with a "Mark this page as patrolled" link at the bottom right.  Snotty Wong   verbalize 20:43, 10 January 2011 (UTC)


 * OK so  far for the bot to put the daily list on the special page here. That's where we have immediate access to  them when we want  to  work  on  them. However, the New unreviewed article banner, because many  of them  will  already  have this banner, and it  disappears as soon as the slightest edit is made to the page even if it is not checked as 'patrolled', may  need a rethink. Remember that putting  a visible sign on the page isn't  going  to help much when everyone has already ignored the page for 30 days. I think there would be a risk that any good faith editor might remove it because s/he might  see from the page history that it isn't really a new page any more. What  we need, I think, is something  more discrete. 'ark this page as patrolled' may  have the wrong  effect too. Some editors may  think that's what  they  have to  do, and just  click on  it  without  doing  any  work  on  the article. We have to  think  of something  idiotensicher - i.e. think  of every  possible thing  that  inexperienced editors get wrong or misinterpret. Kudpung (talk) 14:58, 11 January 2011 (UTC)
 * Unfortunately, when I actually try to patrol form the sandbox list It doesn't work. There is the error "Cannot mark as patrolled / You need to specify a revision to mark as patrolled." (I tried The Who Tour 1973) This patrolling is a flow thing, afterwards you work with buckets such as cats or lists.
 * But I am not really worried about that nor about the tag removal. (Did we already check whether there is a filter that prevents at least the creator from removing it?) Let's do something simple and put it in place and see whether it helps us to maintain the 30 day thing or if we do need to do more about the tail. Actually I fear that something too sophisticated even decreases the incentive to patrol in time.--Tikiwont (talk) 18:01, 11 January 2011 (UTC)
 * You're right, the links aren't working quite right. It's because I have the incorrect rcid in the URL.  I've found a way to fix it and should have it fixed by tomorrow.  The back of the newpages queue has barely been touched in the last few days, so we may actually get a chance to test the bot out in a real world scenario sooner than I thought.  I see your point that actually marking the pages as patrolled isn't terribly important after they've fallen off the newpages queue (and may not even be possible after a certain length of time, I haven't worked out if the rcid's eventually expire).  However, continuing to mark the pages as patrolled is one way to continue tracking the article, since the author of the article can definitely not mark their own articles as patrolled.  There is no mechanism (that I'm aware of) that prevents the author of an article from removing a New unreviewed article template from an article they created.  Snotty Wong   chat 01:07, 12 January 2011 (UTC)
 * And if we're worried about sneaky authors finding ways to reduce the visibility of their articles (i.e. by removing the tracking templates), I can always program the bot to re-check past unpatrolled articles to see if they have been patrolled, and if not to re-add the tracking template. I imagine, however, that we'd probably need to get a wider consensus before we do something like that.  Snotty Wong   spout 01:33, 12 January 2011 (UTC)

Yes, what we need to  do  now is something  that  will  track these pages and provide us with  a basis to  work from, with  a minimum  of fuss, discussion,  and consensus peddling. Of course, if we're later going to  aim  for any  of the solutions I  suggested on  your talk  page, there is going to be endless discussion.(remember getting the BPPROD of the ground?). BTW, I had almost forgotten that I  made some notes and suggestions in  a user page quite a while back. Do take a look, it might be useful. Kudpung (talk) 04:59, 12 January 2011 (UTC)
 * I think your suggestion of posting a notice on the creator's talk page would require wider consensus to implement. And, I kind of disagree with it.  The fact that the article hasn't been patrolled is totally out of the control of the author.  It's not their fault, why should they be getting a notice about it?  In any case, when the backlog catches up, I plan on having the bot add Unreviewed new article to the top of the page (if it isn't already there) and listing the page at the 30-day subpage (hopefully with a link that will successfully allow you to mark it as patrolled).  I say we start with that and see where it goes.  If there is an obvious need for additional actions or a change in the way it works, it will quickly become evident.  Snotty Wong   squeal 06:22, 12 January 2011 (UTC)
 * Of course it would need huge, probably interminable discussion, and the reason why  it's in  my  user space is only  because it's some private notes I  made of some ideas. It's not  up  for discussion  yet. However, I look at it this way: It is the author's fault - s/he's an SPA, doesn't bother to  read any instructions at all although they are visible all round the editing  window, and never comes back  to  take a look. Those are not  the kind of people who  are likeley  to  be coaxed into  becoming  regular Wikipedians. It's like a BLPPROD, if you  don't  catch  them  while they  are sill online and logged in, you  never will,  because most  BLPs are written  by  SPA too. OK, all I'm interested in at the moment as you  know,  is tagging them and having  a record of them somehow, so let's give your ideas a trial spin :) Kudpung (talk) 09:32, 12 January 2011 (UTC)
 * My point is that even a well-written article can fall off the end of the backlog. Just because an article isn't reviewed for 30 days doesn't automatically indicate that it is substandard or that the author neglected the article in some way.  Anyway, the aforementioned URL bug has been fixed.  Now, the links that the bot posts will lead to a properly "patrollable" article (see User:Snotbot/sandbox for an example of how the 30-day list page will look).  I think the bot is ready to go and is now just waiting for the backlog to catch up.  In 2-3 days, we should be able to fire it up and watch it work.  Snotty Wong   gossip 17:51, 12 January 2011 (UTC)
 * Oh yes, there certainly are occasionally pages there that could have been passed as 'patrolled' without needing any tags at all (well, maybe a stub tag, or some additional cats). This often happens because the subject content of such pages is often too hard to understand. Fine, as soon as the list creeps back to 30 days, we can give it whirl. Good work Snotty. Kudpung (talk) 01:52, 13 January 2011 (UTC)

Trials
(Copied over from Snottywong's talk  page so  everyone can  see what  we're doing)

Hi SN. Get ready with  your bot. I don't  know what kind of  time zones are involved but  it looks to  me as if we're back to 30 days already. Cheers, --Kudpung (talk) 10:05, 17 January 2011 (UTC)
 * Will do. I don't think we missed any articles from the 15th or 16th but I'll run the bot for those days in a little while to be sure.  Snotty Wong   express 14:26, 17 January 2011 (UTC)
 * Actually, it looks like there were 54 unpatrolled articles from 12/15 alone... Haven't run it for 12/16 yet. I'd like to supervise the bot carefully for the first time it makes edits in article space, and I don't have the time to do that at the moment.  I'll likely tag the articles and add them to the list in a few hours, unless I find time sooner.  Also, since my trial approval of the bot only covers 50 edits, I won't be able to tag any unpatrolled articles from 12/16 until I get full approval.  Snotty Wong   spill the beans 16:54, 17 January 2011 (UTC)


 * You'll get the apporoval because that's already enough to prove a point. Now you see the importance of what we're doing. In the next 24 hrs, about 90 are going to fall off the yellow pages. I've stopped working at the bottom of the list to give it a chance to build up.Kudpung (talk) 17:09, 17 January 2011 (UTC)
 * Looks like an additional 23 fell off the list on 12/16. Snotty Wong   chat 17:25, 17 January 2011 (UTC)
 * Ok, the bot tagged 53 articles (one of the 54 had already been deleted but not marked as patrolled) and added the list to the 30-day list page. There were a few hiccups that I had to fix manually, but other than that it went pretty smoothly.  I'll wait for bot approval before tagging the ones from 12/16.  If you see anything strange with the articles that were tagged, let me know.  Snotty Wong   spout 23:42, 17 January 2011 (UTC)
 * Checked through everything and all  seems to  be fine  so  far. Press hard for fast  bot  approval  because accordiing  to  the first  run, even at  a conservative guess, it  means that  a massive 18,000 unpatrolled pages per year  are slipping  into  the encyclopedia.
 * There's just one thing  I hadn't thought of: How will anyone working through the list at New pages patrol/30-day list know if a page has been addressed? is there any way  of making  it  yellow like the New Page list and the entries turning  white when the pages have been  detagged? Or is this  too  complicated? Could users just  put a ✅ ✅ template on  the entry, or is is best  not  to  edit  this list  manually  at  all? Kudpung (talk) 07:12, 18 January 2011 (UTC)
 * Bot approved...  I could make the bot highlight the lines yellow while they're unpatrolled, but that would mean that the bot would have to go through all of the articles every day and check them for being patrolled.  Right now, that's not a big deal, but if the list grows to be thousands of articles long, then that could take awhile.  I asked the question above but never really came to a consensus on what the 30-day list page's primary purpose was going to be: for keeping statistics or for assisting patrollers with catching up on the queue.  If it's for statistics, then the page shouldn't be modified.  If it's for assisting with catching up on the queue, then it should be modified.  I keep a record of all unpatrolled articles in text files on my hard drive, so if we need to compile statistics, it will be easier to do it from those files anyway.  But if someone else wants to compile statistics, they could do so from these pages.  I'll take a look and see if I can figure out a more efficient way of checking articles for patrolled status.  What I'm doing right now is actually loading the entire patrol log web page for each article and searching for "No matching items in this log."  There may be a better, faster way to do this through the API.  I'll keep working on it and I'll run the bot for 12/16 and 12/17 in a bit.  <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#FEF7E3;color=#648113">Snotty Wong   spout 14:43, 18 January 2011 (UTC)
 * Whoa, ok, found a much faster way to check for patrolled status. I'll make some changes today.  <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#DDE4C4;color=#648113">Snotty Wong   speak 15:45, 18 January 2011 (UTC)

Technical problem
Well, I did a lot of coding today only to find out that an article can no longer be marked as patrolled after 30 days. Even if you have the correct rcid (recent change id) for the article when it was created, the rcid expires after 30 days. So, tracking the patrolled status of the article on this page is going to be impossible unless we can get the expiration time for an rcid lengthened. I'm going to revert back my changes from today so that the articles are not displayed with the yellow highlight. I'm also going to start a thread on WP:VPT to see if there's any possibility of lengthening the expiration date (which I believe is set using this variable. <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#EEEEFE;color=#648113">Snotty Wong   talk 00:18, 19 January 2011 (UTC)
 * I've started a thread at Wikipedia:Village_pump (technical)#rcid expiration time. <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#DDE4C4;color=#225DC8">Snotty Wong   gab 00:34, 19 January 2011 (UTC)
 * I've left a comment there but I don't think it will help much. It refers to  the discussion  in  October where WSC was simply told that  the NP patrollers should work  harder instead of looking  for  technical  solutions. --Kudpung (talk) 15:35, 19 January 2011 (UTC)
 * Well, at least we can mark the articles as unreviewed and add them to the category, and keep exact stats on how many articles go unpatrolled. It's a start.  We'll see where it goes from here.  <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#DDE4C4;color=#DD0000">Snotty Wong   communicate 16:35, 19 January 2011 (UTC)
 * Precisely. I think  what  you  have now  done is excellent. We now have a handle on  these articles, and something  we can quote at the people who  believe letting 100+ crap  articles loose on the 'pedia every day is no big issue. Our next task will be to encourage the less experienced new page patrollers to  slow down a bit and correctly address the required items on the main  NPP  project  page. A lot of pages get  passed that should not be, and a great many pages are tagged without clicking the mark  this page as patrolled label. Kudpung (talk) 01:39, 20 January 2011 (UTC)
 * We now have bot runs for 15, 16, 17, 18, December. How quickly can we get this up to  date to 20 January even by doing it the same way? Kudpung (talk) 05:32, 20 January 2011 (UTC)
 * It only takes the bot a few minutes to generate the list for each day, however I can only generate the list for articles that are over 30 days old. So, articles created on December 19 are the newest articles that I can test at the moment.  The list is just about as up to date as it can get.  <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#EEEEFE;color=#00AA00">Snotty Wong   verbalize 06:33, 20 January 2011 (UTC)
 * Yes of course - silly  me! What  must I have been thinking  about? Too  much  stress over poor NP patrolling  probably  ;) Kudpung (talk) 07:04, 20 January 2011 (UTC)

Snotty, thanks for implementing my idea to highlight the pages in yellow. This project is now doing  an excellent job,  and I/we can now perhaps consider sending  a newsletter, when we have a month's worth  of trials, to  all  users who  are subscribed to  the NPP project, and those in  the category  who  are sporting  an NPP userbox (the lists are not identical). Kudpung (talk) 14:21, 25 January 2011 (UTC)
 * I like the idea of sending out an announcement in the near future, perhaps in a week or two. I've been thinking of a few changes that should be made:  Firstly, we're going to need to create a subpage for each month.  With 25-100 articles per day, we could end up with 3000+ entries in a month.  We're only 1.5 weeks in with 425 articles and the page is already 127kb.  Secondly, I think we should move the page to a more descriptive title.  "30-day list" makes sense to you and me, but may not be relevant to those unfamiliar with the concept.  I'm thinking maybe WP:New pages patrol/Unpatrolled articles/December 2010 and WP:New pages patrol/Unpatrolled articles/January 2011 and so on.  And the root page, WP:New pages patrol/Unpatrolled articles could serve as an index page for all of the monthly subpages.  Any thoughts on that?  <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#E0F4FE;color=#DD0000">Snotty Wong   chat 15:05, 25 January 2011 (UTC)
 * Sounds logical, yes I'm definitely  in  favour so perhaps you  could go ahead and make the changes. Perhaps we should move this thread, Back of the unpatrolled backlog, to the talk  page of the root page too,  and leave this talk  page for  general  banter about  NPP. As regards making  the project  known, we have several  options: the VP, a newsletter to  all known patrollers, and the Signpost. WE can aso make a teplate that  a bot  can deliver to  all the Wikipedia projects and sub-projects. There may be some other channels too, but  I'm not  sure what they  would be. I can include a mention of it in the WP:SCH newsletter that  I've written and been meaning  to  send out  for nearly  2 months. All this may  have the effect of increasing  NPP and doing  it  well, which of course is what we ultimately  want,  the purpose of our 30-day  exercise being basically to  provide the proof that in  practice, NPP is far from  perfect. Kudpung (talk) 04:23, 26 January 2011 (UTC)
 * Pages moved, talk page thread moved. <span style="font:13px 'Copperplate Gothic Light';border:#AAAACC 1px inset;background-color:#F2F9FA;color=#25900D">Snotty Wong   yak 18:15, 26 January 2011 (UTC)
 * FWIW, I think this division by month was a truly excellent thing to do, something we should have done much earlier. Normally, a change like this needs to be widely discussed, but it is such an obviously desirable thing that I would suggest first announcing it, at both the VP and ANB, and, if there are no objections, making it a permanent practice.     DGG ( talk ) 04:39, 28 February 2011 (UTC)
 * Well, I think we should hold off on announcing anything until the backlog at Special:Newpages becomes a problem again. Somehow, either through a technical glitch or an overzealous patroller, the entire 30 days of backlog got mysteriously cleared and we're down to a few hours of backlog on average (currently at about 16 hours).  I'm probably going to shut down the bot for awhile until the backlog approaches 30 days again.  I'll run it manually every once in awhile just to update the patrolled status (i.e. yellow highlights) of the existing articles in the list.  <b style="white-space:nowrap;text-shadow:#000 0em 0em 0.4em,#a00 -0.2em -0.2em 0.4em,#00a 0.2em 0.2em 0.4em;color:#ddd">&mdash;SW&mdash;</b> comment 14:59, 28 February 2011 (UTC)
 * We need to establish technical glitch or an overzealous patroller. Whichever it was, it would, on the face of it, be most unusual for what has been a monthly average of about 6,000 articles/30 days over the last 6 months to suddenly dwindle to 50 articles/10 hours. Shutting down the bot for a while is a good idea if there is nothing over 30 days. Kudpung (talk) 16:44, 28 February 2011 (UTC)

Now seems to be just less then a day now that is unpatrolled
Okay lately it seems so many articles have been patrolled that sometimes a page that has not yet been marked for 24 hours is gone. This is getting kind of annoying a way, since now a page can be never seen then. Which now makes me wonder-what happens to the pages that never get marked as patrolled that are less then a month old that go away? (A week ago there were still some January pages, now its just the past 24 hours!)Kamkek (talk) 01:40, 18 February 2011 (UTC)


 * Yes, there appears to be a problem. The Special:NewPages seems to  be listing  only  50  articles. The older list  page won't  load and it  it  not  possible to  get  to  the back of the list. We're looing  into  it. --Kudpung (talk) 13:29, 21 February 2011 (UTC)

Okay thanks, yesterday afternoon there were only 4 listed! (Granted they disapeared fast since there were not that many then)Kamkek (talk) 16:04, 21 February 2011 (UTC)
 * I've asked WSC on this, and I believe he asked a Q on the VP tech. Are we getting any  feedback on this yet? Kudpung (talk) 15:00, 27 February 2011 (UTC)
 * Catching up. It  seems as if the discussion is here. Concerned editors and NP patrollers please check it  out. Thanks. Kudpung (talk) 15:21, 27 February 2011 (UTC)

AS of today, the last five at  the bottom of the list: ({{age|21|05|1959}})| |place of birth = Hermosillo, Kudpung (talk) 18:30, 4 March 2011 (UTC)
 * 17:56, 3 March 2011 ‎Pepsi-Change the Game ads (hist) ‎[601 bytes] ‎Aditya Simha (talk | contribs | block) (←Created page with ' '''Pepsi Change the Gam... {')
 * 17:35, 3 March 2011 ‎Ray Donn (hist) ‎[3,133 bytes] ‎Valleyman1970 (talk | contribs | block) (←Created page with ' { {')
 * 10:03, 3 March 2011 ‎Jesús Ochoa Domínguez (hist) ‎[4,840 bytes] ‎Mobset1980 (talk | contribs | block) (←Created page with ' {{{Ficha de actor |bgcolour = |name = Jesús Ochoa Dominguez |date of birth= May 21 1959
 * 03:49, 3 March 2011 ‎Ellie Rose (hist) ‎[1,140 bytes] ‎Pghn (talk | contribs | block) (GOOD page.) (Tag: possibly non-minor edit)
 * 02:01, 18 February 2011 ‎Till the World Ends (hist) ‎[3,666 bytes] ‎Nickyp88 (talk | contribs | block) (←Redirected page to Femme Fatale (Britney Spears album))

Back to 30-days
The backlog has now reached 30 days again having caught up  from  the hard work that  was done to  clear it in  January. Emphasis is nevertheless on quality rather than speed of tagging - poor NPP work (both tagging and not tagging) is the cause of many  peripheral problems that  may  not  seem evident to  everyone here. The Snotbot has now been relaunched to keep these 30+ day  pages on  the radar. --Kudpung กุดผึ้ง (talk) 07:58, 27 May 2011 (UTC)

Tagging articles
Hi, I tried to get a brief picture of your project, but I fail to understand some things. If it is not absolutely necessary to tag all unpatrolled articles, can you please just limit the bot to list the unpatrolled articles in your project lists. It is confusing to see these tags appear on the articles and time consuming as one starts looking around to see where they come from, what they are all about etc, when IMO this is a project task. Thank you. Hoverfish Talk 17:35, 30 May 2011 (UTC)


 * If an article is left unpatrolled, it  is because no  patroller has  been able to  apply  his or her knowledge to to  what  needs to  be done to it, and has correctly  left  it  to  someone else. Unfortunately, someone else is usually  also  patrolling  the new articles as they  arrive, and as soon  as the articles slip  off the live feed, they  rely  upon being  tagged or passed by someone working  laboriously  through  Special:new pages. The deeper those articles sink towards the back  of the list, the less attention  they  attract, until editors with  more experience have a crack  at  the backlog. Articles must  either be passed as OK by  clicking  the 'mark  this article as patrolled' button, or tagged for attention,  and in  the worst  case scenario, tagged for one of the methods of deletion. Anyone dong  new page patrol is part of the Wikipedia community and hence part of the the project, whether they  are wearing  a New Page Patroller userbox on their user page or not. If  the patrollers have not  been able to  fix the problems immediately,  the tags they  place will  draw the attention  of other editors who  may  be able to  fix them. If on  the other hand there are articles that  are tagged but  are still  highlighted in  yellow, this is a software glitch, and please let us know about it.  --Kudpung กุดผึ้ง (talk) 15:58, 31 May 2011 (UTC)


 * I saw a lot of articles that were tagged with "unreferenced" or "uncategorized" and they were still appearing in yellow. I marked as "patrolled", because someone had tagged them, but it seems like a software glitch, if I understand you correctly. Divide et Impera (talk) 20:18, 10 June 2011 (UTC)


 * Theoretically (with one or two exceptions) if an article is tagged and/or edited, it is automatically marked as patrolled and shoul no longer appear highlighted in yellow. Th workaround is to manually click the 'mark this page as patrolled' link. --Kudpung กุดผึ้ง (talk) 14:28, 11 June 2011 (UTC)


 * And that's what I did: I clicked on the 'mark this page as patrolled' link. Only that it wasn't one or two exceptions, it was all of the yellow ones, which even if tagged, had to be clicked. Divide et Impera (talk) 14:31, 11 June 2011 (UTC)


 * OK. We'll try  to  find out  who  the devs are that  can  fix this glitch. The person who originally  designed the operation  of Special:new pages no longer edits Wikipedia and cannot  be contacted -  no one else apparently has the source code. Nevertheless, the ones that  are not  tagged should of course only  be passed as acceptable if they  do  not  need any  attention at all. Kudpung กุดผึ้ง (talk) 00:38, 12 June 2011 (UTC)

Improvements to NPP
A discussion on  the improvement  of NPP  is taking  place at meta Wiki. Kudpung กุดผึ้ง (talk) 02:18, 1 July 2011 (UTC)

Is this project still active?
I see nothing since last September. Could someone give an update on the project or mark it as historic/inactive? thanks -- Michael Scott Cuthbert (talk) 23:32, 1 July 2012 (UTC)
 * It is very much active, and although it has taken over a year to get it off the ground, the prototype of a completely redesigned new page feed is available for testing  at Special:NewPagesFeed. Do try it if you can, and provide some feedback. Kudpung กุดผึ้ง (talk) 01:19, 2 July 2012 (UTC)
 * Looks great; fast and easy to use. Is there a difference between Patrolling a page and marking it as Reviewed?  I'm going back from the earliest unreviewed and there are some no-brainers for patrol, but not sure if review has a higher standard.  I've done some very obvious ones but I want to be sure that the standard isn't different before marking any I would patrol, but which barely pass.  Thanks! -- Michael Scott Cuthbert (talk) 02:23, 2 July 2012 (UTC)