Some good ideas here. They will only ever be of any use if New Page Patroller is made into a user right. The ideas here do not address the serious and immediate problems, such as for example, the fact that New Page Patroll attract the newest, least mature, and least experienced of all users who do not understand the principles involved, and who do ot read the instructions. A solution to this immediate problem was proposed that would restric the creation of new pages to editors. The restriction was adopted in Mrch 2011 by a clear majority consensus on a heavily subscribed RfC, but the request to implement the required software change was refused by the WMF. Kudpung 07:16, 20 September 2011 (UTC)
Topic on Talk:Page Curation
They will only ever be of any use if New Page Patroller is made into a user right. The ideas here do not address the serious and immediate problems, such as for example, the fact that New Page Patroll attract the newest, least mature, and least experienced of all users who do not understand the principles involved, and who do ot read the instructions.
If the quality of NPP work by new patrollers remains low even with an improved interface (and instructions for new patrollers can be simplified and made more accessible), there may be additional options as well to train new folks on NPP. For example, it would be possible to allow any registered user to perform proposed NPP actions, but to require sign-off by an experienced patroller, who could leave quick feedback for the new patroller and ultimately promote the user to have the same permission-set. In that way, the experienced NPP pool could scale significantly faster than through more conventional user right promotion processes relying exclusively on an editor's prior track record.
We don't have an 'experienced NPP pool'. We have - according to the research and stats that has already carried out since October 2010 by Snottywong, Blade, and myself - a vast army of occasional, extremely inexperienced patrollers who appear to ignore all attempts to educate them, and a just a handful of experienced editors much of whose work is cleaning up after the other patrollers. Any solutions concerning improvements to NPP would require software implementation to prevent just anyone from patrolling and tagging pages. Solutions are really needed now, not after another year of discussion by people new to the problems and who don't understand them.
Well, that might be an exaggeration. I think there are more than just the three of us experienced users patrolling (and I have been quite inactive at NPP lately, since I've been exploring AfC). At least at the beginning of the year, there were some more experienced users patrolling. See these stats. Snottywong 15:33, 20 September 2011 (UTC)
I think he was saying that the three of us carried out the study, not that we were the only three (although it sometimes feels like it). Perhaps unsurprisingly, I don't think a change to the interface will materially change the quality of NPP. As I have said elsewhere, nothing short of a hard restriction (this applies both to article creation and to being allowed to do NPP) will get anything done; people don't read the directions we already give them. An improved interface won't prevent someone from writing "Let's expose all burakus for what they are!!!!!!!!" with a link to a list of burakumin in Nagano Prefecture, nor will it stop a 13 year old who isn't paying attention from patrolling it; in fact, it will further enable people to mark such blatant attack pages patrolled without checking. We need ways to slow it down, and the only good way to slow NPP down is to slow down the number of pages being created. If it is slowed down, people will be able to check everything more carefully; as it is now, pages like the one described above are slipping through (the example came dangerously close, and if it had gotten through could have done horrific damage to thousands of people). A better interface would be good for patrollers like the three of us, who know what we're doing, but not for those who don't, which is around 95% of the rest of the people doing NPP now.
I agree with you on most of your points. More time spent with each article (in creation and patrol, frankly) will certainly improve the overall quality of the encyclopedia. While efficiency improvements are a focus (ideally leading to more time spent reading and understanding the content and less time interacting with UI), I don't believe that's the primary goal of the proposed interface. Rather, they help to support the primary goals. If I had to outline them as I see them now, I think I'd say:
- Make NPP less complex to learn and vexing to do, to attract more experienced editors into it.
- Create an interface that can be used to train some of those newbies, in order to get more people working on the problem.
The eventual point is to get more (trained) people looking, so each of them can spend more time and be more accurate and less stressed (ie. reduced workload per patroller). Jorm's mock-up is just the first step in that direction, but I think it's a good start.
There's a long history of websites having content moderation problems, and approaching those problems by recruiting more moderators. Websites that are infamous for their trolling have managed to solve this problem without even disabling anonymous contribution, and have been doing so for over a decade. Most such sites have a system for giving moderation privileges only to people who are deemed trustworthy (via some impartial metric). I think such a metric is a reasonable thing for us to have, but we have to address these problems in the right order so as to avoid overwhelming anyone. The best system is probably one that requires agreement between multiple patrollers, since that seeks impartiality as well as accuracy, but (as has been pointed out) that's not possible until there's a lot more patrollers.
There are a lot of existing solutions out there, parts of which are probably worth adopting. Our challenges are certainly unique in some ways, but I don't think they're insurmountable by any means.
I don't think we can Make NPP less complex to learn and vexing to do. Snottywong and I have tried already by recasting the New Page Patrol project page, but short of dressing it with an attractive layout with pretty graphics and buttons - which we have not done, notability & CSD are such extremely complex and ambivalent areas that even after gnashig our teeth and banging our heads against the wall over it, we have not been able to come up with something simpler. Other ideas and suggestions are more than welcome.
I see where you're coming from. Policy is complicated, and it needs to be since the world that Wikipedia tries to model is a complicated place. At the same time, getting people to take the time to understand it is hard.
Let's say each person comes to NPP with a finite amount of time and effort they're willing to invest. Maybe 1-2% of those people manage to become involved enough to understand the entire process and become experienced NPPers. If we want to increase that number, we can take a few approaches:
- Reduce the time and effort required to learn the necessary information.
- Increase the amount of time people are willing to invest, by making NPP feel more rewarding overall, or by making it feel more rewarding further back on the learning curve.
- Increase specialization, such that each individual needs to learn fewer things to be effective.
You've made inroads here, for sure, but we can probably do even better if we put the information directly in the software and present it incrementally as people work. Presumably some number of people do NPP because they find the process of doing it well, the impact it has, or learning in general to be intrinsically rewarding. Those are the people who we want to encourage. If we can create a way for people to have those feelings with less initial investment, they'll be willing to continue making an effort.
Looking at it another way, what we have now is a an excellent textbook on how to do NPP. Let's figure out how to teach the class.
We've been slowly turning the page at WP:NPP on the sly into a tutorial, and I've already done some NPP coaching. Ironically, I'm a teacher and a Grade school and High School textbook author, but let's not run away with the idea that we can sit 600 NPP students in a lecture theatre for a couple of 3-hour sessions, make them sit an MCQ for an hour, and then give them a ticket to tag.
However, I had the idea at the weekend that perhaps I could develop this idea of a screencast one very big stage further, and and make a video tutorial for NPP. Now, that might be a way of at least being sure that they would 'read' the instructions. There's more to this than meets the eye, because this would pave the way to solving another very disturbing problem that I'm working on (Philippe know about it). Nevertheless, all solutions for NPP whatever they are, still look as if we're not going to escape the stick 'n carrot of making it a user right, and at the same time, we must develop some ideas to boost new-user retention.
What I need right now is to find a screencast package that works on my platform.
But Kudpung, if NPP does become a userright, what is to stop an incompetent patroller from slapping db-tags on a page without patrolling it?
I'm not sure I agree with all the above points, but I can't come up with any alternatives right now. Σ 01:03, 21 September 2011 (UTC)
It will allow admins to filter out the people who shouldn't be doing NPP and/or revoke their rights if they screw up too much. If they're flaunting the spirit of the rule, it's rather easier to sanction them for it; as is, the situation now is pretty much like what he described.
I'm very leery of creating a Priesthood of Article Gatekeepers. User rights are not things that should simply be created just because they can be. I'm also opposed to making the process more complicated - which is what additional rights always do.
There has to be some way of preventing people from patrolling if they are doing a terrible job of patrolling. Without a user right, the only other alternative is to block them, which is non-ideal. And making a user right might have the added benefit of creating a (likely temporary) increase in the number of patrollers, since there are a lot of user right whores out there who will be clambering for the latest badge... ;)
So, I get what you're saying. It's clearly a problem, and probably a super-frustrating one. Here's an idea I've been kicking around:
Maybe we have a user right - pagepatroller - but instead of it being granted to you by a Magical Wizard, you'll automatically earn it over time. It can still be taken away from you, though - which will set you back to step 1.
Let's say it works like this:
Bob decides to page patrol. He fires up the interface and goes to town. His first attempts at page patrol, however, automatically require review by someone who has the pagepatroller right. Say, his first ten. He patrols ten pages, and they're marked as "patrolled, but needing review" (so technically unpatrolled).
Then, someone who has the right (say, Snottywong), can review his reviews. And mark them as "good" or "bad". Once Bob has ten reviews marked "good", he is automatically awarded the pagepatroller user right (and with it the ability to review other people's patrols that "need review").
Obviously the numbers will need working out, and it will increase workload in the very short term (but possibly not, since, as you say, you guys are already re-reviewing anyway), but will decrease it overall (since you won't have to re-review these guys once they earn the right).
And if Bob goofs up? A conversation and a couple clicks by an admin and the right is revoked. He has to start over again, and earn ten "good" re-reviews.
What I'm trying to avoid is a situation where there's a whole level of bureaucracy that a potential patroller has to go through before they can start (another "RfA" type process). We should be open with letting people into the door, but be able to push them back out if need be.
Obviously, the numbers involved can be modified. And with decent messaging we can make this user-acceptable and friendly:
"Hey! You've just patrolled your first page! That's great! Just so you know, everyone who starts patrolling has a grace period where their work is reviewed by a more experienced patroller. Once the review period is over, we'll let you know, and you'll then be able to do reviews yourself!" or something along those lines.
What do you think?
I feel like this is becoming unnecessarily complicated. I'm certainly not envisioning an "RfA" type process for handing out pagepatroller user rights. That would be way over the top. I think it should be handled like all other uncontroversial user rights are handled, like w:Wikipedia:Requests for permissions/Rollback or w:Wikipedia:Requests for permissions/Confirmed. We would just set up another Request for permissions subpage like w:Wikipedia:Requests for permissions/Pagepatroller, and anyone can ask for the permission there. As long as they satisfy a basic list of criteria, then the user right would be given to them. As for the criteria, I'm envisioning something like:
- Account has more than 500 edits
- Account is more than 3 months old
- If pagepatroller user right has been revoked in the past, require an explanation for how previous problems have been addressed
The numbers could be tweaked, those are just off-the-cuff suggestions. If a user satisifies the above criteria and has enough clue to actually find the place to request permissions, then there's a good chance they have enough clue to patrol pages at a basic level. This doesn't mean that we can't or shouldn't still build in some kind of process to allow review of their patrols by more experienced users, but I think handing out the user right should be handled like all other uncontroversial user rights are handled.
One of the principle reasons for advocating NPP as a user right is that the only recourse we have at the moment for patrollers who persistently refuse to improve is to block them for being disruptive to the process. We have sadly had to do this on occasions, but it's a bit drastic, like taking a sledge hammer to crack a peanut. We should be looking at ways to retain potential good (and perhaps young) new editors rather than piss them off entirely. As SW says, a user right can easily be revoked without much ado. Whether user rights, like adminship, are a big deal or not, they give people a sense of responsibility and the desire to do a better job.
This forum software is so confusing :( It appears that I have made the very same suggestion as you somewhere else on this page today. Would just add to that that there should be a guideline for the admins who accord the rights. Naturally they would seriously need to check the user's talk page, and not accord the right to anyone who has been warned or blocked for anything, or who has received CSD, PROD, or AfD notices for their own creations, and whose own creations are tagged, or missing cats, stubs, refs, and anything else that NPPers themselves are supposed to repair on other people's articles.
SW, I couldn't agree more. So I've expanded a bit on that theme in another thread (I think) on this forum - if you can figure out how this forum works, because I cant!
Serious editors and admins at en.Wiki realised from their solid year of research and experience that turning the vast army of inexperienced and ineducable NPPers into gatekeepers would probably not be feasible, so they agree with you there, if not for the same reasons. So they looked at another solution - which you also summarily, and apparently personally, dismissed.