Wikipedia talk:Autoconfirmed article creation trial/Archive 3

src: farm3.staticflickr.com

Video Wikipedia talk:Autoconfirmed article creation trial/Archive 3

Proposal halaman arahan

@Kaldar and DannyH (WMF): , I'm sure you're tired of me as a pest now, but does WMF have a graphics team that can update the flow/image of the landing page to be more current? Or is there a picture in Commons that we can use. I really think that giving the page a refresher in terms of visual appeal will make it easier for new users to engage on Wikipedia. TonyBallioni (talk) 20:50, July 26, 2017 (UTC)

Hi TonyBallioni, that's a good idea. We do not have designers on the Community Tech team, but I'll ask one of the designers on the other team if he has time to see it. I will tell you what he says... - DannyH (WMF) (talk) 21:34, July 26, 2017 (UTC)

Yes, I know how departments are often stretched, but I think the first page that many new users will see is the PR type for the Foundation. To be honest, I would be happy if the Foundation's communication team was involved in the process here to help us deliver a good welcome message to people. TonyBallioni (talk) 02:02, July 27, 2017 (UTC)

• I am now writing this some other place, but this is the place to say it. The landing page should make this process look completely normal, like a driver ed or like a new employee orientation in a company. Very friendly, very normal - "Welcome to WP! Working in WP is so complicated that we help new editors create their first few articles with a guided process, and then review them before their article goes to the free encyclopedia. Something like that. Jytdog (talk) 03:13, August 19, 2017 (UTC)
  • Agreed. A clear and positive process for learning how to edit/write WP. Doc James (talk Ã, Â · contribs Ã, Â · email) 06:54, 20 August 2017 (UTC)

• If you're going to create a new article, we've put some outbound features to help you through the process and it will be seen by Wikipedia visitors as soon as it meets our minimum standards for display. Kudpung ??????? (talk) 07:56, 21 August 2017 (UTC)

Jytdog, DocJames, Kudpung: The current mockup for the landing page is here: Wikipedia: Authenticated article creation experiment/Mock-up landing page. That's the text we plan to use. Let me know if you want to make changes to the text - we really can do it, we just need to know what you want to say. Ã, - DannyH (WMF) (talk) 19:08, 21 August 2017 (UTC )

• Thank you, that looks good to me. It advises people to read Wikipedia: Your_first_article and that talks about direct article creation. User: Kudpung and User: TonyBallioni is there any plan to change the guide document for the trial? I would love to do it if you like... Jytdog (talk) 20:16, 21 August 2017 (UTC)

------------------------------------ - ------------------------------------------------- - ------------- JytdogTonyBallioni , ironically WP: YFA is the first thing that scares the new user away. That is exactly the opposite of the intended effect. We know 40 years ago from our research at TU Berlin that there are two possible reactions to such a text wall (now in the context of the 21st century) Ã,:

1. Sheesh! if this is what I have to do or know, I will not bother. I will put my article in WordPress instead.
2. This attack is for birds! I will write my nonsense pages. No one knows who I am.

WP: YFA has been on the 'to do' list for years. Full rewriting is required, condensation up to one third of TL; DR current, and a very strong emphasis to use the Wizard. I'll do it, but by looking at what's happening, and some rewriting the Wizard page so that non-genuine users (50% of new accounts) can understand it, there's a limit to what I can do in the remaining time. I will help with that. Kudpung ??????? (talk) 22:25, 21 August 2017 (UTC)

• Respond to the Kudpung ping in this section as it relates to graphics: I know Drewmutt has some thoughts about graphs in the past: he might be able to give me better feedback than me. I prefer something more contemporary and it does not look like a clipart from 2006 (do not take offense with Puzzly, but he reminds me a lot of Office Paperclip). TonyBallioni (talk) 02:45, 22 August 2017 (UTC)

YFA

• FWIW, JytdogTonyBallioni , I have done WP: YFA now. It's not as short as I want but at least someone from South Asia will understand it now - if they bother to read it. Kudpung ??????? (talk) 04:01, August 22, 2017 (UTC)

Kudpung, I see WP: YFA to see the edits you made, but I do not see anything in history. Are you publishing your new version? - DannyH (WMF) (talk) 16:55, August 22, 2017 (UTC)

No. It's on the user's sub page. Kudpung ??????? (talk) 18:22, August 22, 2017 (UTC)

Oh, I see. I think the DO and DO NOT parts in your version are a bit confusing - in the DO NOT list, items are usually things that should not be done by people - DO NOT write articles about yourself, DO NOT repeat existing articles, etc.. In your list, it looks like you said DO NOT consider registering an account, DO NOT find Wikipedia first, DO NOT practice first. - DannyH (WMF) (talk) 20:17, 22 August 2017 (UTC)

What is it? Kudpung ??????? (Talk) 23:56, August 22, 2017 (UTC)

Yeah right, is not it? It looks like my last save failed to load - it's happening here. One of the problems is that 9t's suc a WoT is hard to know from where to start. Maybe Jytdog wants to go. Kudpung ??????? (talk) 00:47, August 23, 2017 (UTC)

• I took a shot just simplifying the beginning part that probably everyone will read... diff. Mind? Jytdog (talk) 02:40, August 23, 2017 (UTC)

Text of the landing page

@Kal and DannyH (WMF): I have made some minor changes. Nothing can not be fixed. I have no opinion about the graph. Maybe Tony has an idea, there are several million pictures in commons to search... Kudpung ??????? (talk) 22:48, 21 August 2017 (UTC)

Kudpung, you add the phrase "it will be seen by visitors to Wikipedia as soon as it meets our minimum standards for display" to the landing page. Is there a description somewhere about that minimum standard? This is a vague phrase, and I think it is potentially confusing, if we do not have something specific to link to. Also ping Jytdog and TonyBallioni. - DannyH (WMF) (talk) 15:14, August 22, 2017 (UTC)

They will be WP: N, WP: NOT, and WP: BLP. Not sure how you would process it into text. TonyBallioni (talk) 15:48, August 22, 2017 (UTC)

Do not forget WP: V Legacypac (talk) 16:03, 22 August 2017 (UTC)

Hmm. It's very hard to figure out what to put in, which is why every attempt to intro to new people ends up as a "wall of text".!, - DannyH (WMF) 16:52, 22 August 2017 (UTC) g

The whole issue of Wikipedia is that it puts people with rules, rules, acronyms, and steep learning curve right in their face at the beginning. It why new users are not staying. Unclear is the best at that stage - after all it's just the opening page. The wizard will confront them will text the wall immediately. Kudpung ??????? (talk) 18:12, August 22, 2017 (UTC)

Maps Wikipedia talk:Autoconfirmed article creation trial/Archive 3

Suggestions for upgrading article wizard

There are several suggestions for improving the article wizard on Wikipedia talk: a wizard article that might be of interest. thanks. Darylgolden ^{( talk )} Ping when replying 06:57, August 27, 2017 (UTC)

src: www.mariowiki.com

Initial analysis of patrol and article creation workload

Updated August 26, 2017: I've collected better data on article creation after I found our previous data excluding deleted articles. This affects the analysis for findings # 1 and # 2 below, but the conclusion in finding # 2 has not changed. Updated August 30 & amp; 31, 2017: I've collected better data on the number of patrols and active patrols after being notified that we did not include patrols that were performed outside the PageTriage extension. All charts have been updated to reflect the new data source. This update primarily affects the number of active patrols as we find many alternative tools for patrol users. I've updated a quick analysis of each graph to reflect the current state of the graph.

Hello all! Over the last few days I've done a preliminary analysis of data about workload and article-writing patrol, and found some interesting results I want to share with you.

Quick background: We are interested to understand the effect of ACTRIAL on the Wikipedia quality assurance process, and as part of it has proposed a hypothesis on how this will affect the New Page Patrol. Referring to our project page in meta, we would like to know more about the following hypotheses:

• H9: The number of patrols will be reduced.
• H10: The number of active patrols will be reduced.
• H11: Distribution of patrol activity is balanced.

We also have two related steps for H9 and H10, see how the number of active patrol and patrol actions is related to the number of articles created. I will summarize the findings that I find interesting. If you are interested in more details, I describe our method on the project page on meta and there are larger notes and graphs in my work notes from August 22 and August 23. Since this is a preliminary study, the focus is on general trends, and I will try to explain what I see as we walk.

Finding # 1: The number of articles created per day is generally stable. The graph above shows the number of articles created from 1 January 2011 to 1 July 2017, as well as draft publications from User and Draft namespaces. Since the second half of 2014, the number of articles made per day has been fairly stable and averaging about 1,100. The use of draft space is stable, although since the beginning of 2017 there have been many publications of the draft Namespace. From what I can find this is because AfC works to reduce their backlog.

Find # 2: New Page Patrol does not seem to meet the request. The graph above shows the proportion of patrol actions performed against the number of articles created and published starting October 1, 2012 (the first full month after the PageTriage extension was introduced) until July 1, 2017. This proportion is rarely above 100%, meaning that the backlog generally increased.

Find # 3: The number of active patrols used increases. The graph above shows the number of active patrols per day from October 1, 2012 to July 1, 2017. Back in mid-2013, there appears to be around 100 active patrols every day, and the number continues to rise until early 2015. Activity increases again at the end 2016, possibly due to future recognition of new PageRp Rights users. Once NPR rights are introduced, the number of active patrols per day is stable around 75.

Finding # 4: Most patrol jobs are performed by the most active patrollers. The graph above shows the proportion of patrol actions performed by 25% of the most active patrollers. Generally this makes up 70-90% of all patrol actions. We saw earlier that there are currently about 75 active patrols per day, which means that about twenty patrollers do most of the work. I think it raises concerns about the extent to which patrols work too much and burn, but also something that might motivate NPPers to recruit more patrols and encourage inactive patrols to participate more.

Added a note here that I've updated the graphs for Findings # 1 and # 2 after I found that our data set does not include deleted articles, new graphics including deleted articles. The general trend at # 1 is somewhat different, the conclusion in # 2 is the same. I've added a note at the top as well. Cheers, Nettrom (talk) 23:14, August 26, 2017 (UTC)

2017 The uptick on the pages of the draft and userspace also reflects the urge to clear non-afc drafts and user space spaces that continue to appear on many pages Legacypac 23:20, August 26, 2017 (UTC)

Legacypac: I read some archive of AfC talk page and got the impression that some sort of push cleanup is going on. Thanks for confirming that's the problem, appreciate the insights! Cheers, Nettrom (talk) 14:54, 27 August 2017 (UTC)

Yes AfC has pushed backlogs that are mostly ordinary. I'm not sure that generates an uptick on the page from Drafts & amp; User space. I am referring a reduction of 8-10,000 pages here since November ?? 2015 [1] and a reduction of 6,000 net pages in a non-afc design here [2]. Both of these backlogs have seen a large net reduction, although new pages are added daily to the backlog. With the expansion of G13 to All Drafpace a few days ago, we quickly pushed the MusicBot report to almost zero. IE removes, fixes and promotes, assigns, or sends to AFC and hopes for repairs. Legacypac (talk) 17:05, August 27, 2017 (UTC)

Comment on graph

Hi Nettrom, I now have a chance to check this graph more closely. I think they would be easier to read if they had been presented as a line graph (we're not all statisticians) and with a sample period starting at the point that we, the volunteer community, are interested in such a dramatic sharp rise in early 2016 from 6,000 to 22,000, and a sudden lack of interest by patrollers as shown in February 2017 - where no theory emerged.

• Searching # 1: The number of articles created per day is usually stable: How do we explain the marked increase in drafts in Q1 2017 that suddenly fall again? Does one reviewer review do a lot of reviews? Legacypac has provided some possible explanations, but since AFC and NPR can not be separated closely and therefore more importantly, we need some factual data about the AFC project and its operations.
• Find # 2: New Page Patrol does not seem to meet the request. The 2016/2017 figures seem to correlate with the graph I'm generating. We still have no explanation for a sudden change in the mood of patrols - especially in February 2017. It would seem that from the beginning Q2 the percentage is on the rise.
• Find # 3: The number of active patrols is slowly increasing. I see an improvement beginning around early Q4 2016, but the general trend from mid-2017 seems to be going down a bit.
• Finding # 4: Most patrol jobs are performed by the most active patrollers. It seems to more or less reflect what we have built empirically. However it does not provide any substance for the burn claim. I would argue that many applicants for NPR rights have become hat collectors. Of course from being one of the most active admins to fulfilling these rights, I find that many applicants are not new users but have never conducted active patrols before Nov 2016 and have no intention of patrolling regularly, therefore they are entitled to numerical criteria and trouble-free history. Regular reviewers do not suffer from burning because they still do it. This is why these charts will be more helpful if they show shorter sample periods for our current purpose and should be displayed as a starting line in June 2016 2016 which will then show more clearly what happens when NPR introduced in Nov 2016 and trends since.

But I understand very well that WMF, as non-patrol, wants these statistics for their important purpose, while the requirements of the community are very different. We need to look at the end of ACTRIAL how we can directly address and motivate New Patrol's idea, and with user feedback, as we have done on Wikipedia: The Kuration Improvement/Suggestions page (which came a year ago still) received little or no there is official recognition from WMF), to know what we want WMF technicians to develop for us.

------------------------------------ - ------------------------------------------------- - ------------- DannyH (WMF) I do not want you to assume that this is a personal request from me. I have received many comments outside the Wiki that the graphs do not really help us here. What we need is a trend line over a more recent sample period, such as from as early as 2016 and that clearly depicts the moods of the reviewers.

In 2009 or 2010 when WereSpielChequers and I first realized that it became difficult to hold backlog, some intake of more than 3,000 a day. The articles of the day were, of course, easier to patrol - most of them still about traditional encyclopedic topics and the rest was easily recognizable waste. The curious policy of including any soccer player who has ever kicked the ball has not started yet, the penetration of the Internet does not significantly reach the non-English L1 region in the development process, and not so many people have found how easy it is to exploit our unpaid work into making money quickly or fast money.

By early 2012 that number is already half, in 2013 this is the third, and now floating on what seems to be an average of about 900. None of this we do. The crap that we can today should be seen to be believed, but the paid staff will not do it or will not listen. They will not even notice until we recently announced our intention to launch their own ACTRIAL, at which point they entered the left stage in a panic with Wikipedia: New page patrol/Analysis and proposals.

Can we ask members of WikiProject Statistics if they are interested in helping data analysis? I know a little, but I'm not a statistician. I'm glad we've collected data to establish basic data. It will prove to be invaluable later. Mduvekot (talk) 04:41, August 29, 2017 (UTC)

Mduvekot: Yes, anyone interested; it would be very nice. If you'd like a link to skip, meta: Research: Autoconfirmed autoconfigured experiments have all the research questions, and Morten keeps updating with initial measurements and data. We would like to get more feedback and suggestions on how to measure the impact of trials. - DannyH (WMF) (talk) 05:03, August 29, 2017 (UTC)

Graph from 2016 onwards with trend line

I have created a graph showing the period from January 1, 2016 to July 1, 2017, and added the Loess trend line to them. All these graphs show daily measurements.

Articles are created per day, by type
The plot shows the number of articles created per day, divided into three: articles created directly in the main namespace, the page moved to the main from the User namespace, and the page moved to the main from the Draft namespace.

Number of articles created per day
This plot shows the total number of articles created on or moved to the main namespace (the sum of the three measurements in the first graph).

Proportion of patrolling actions to articles created
This plot shows the proportion of patrol actions (either recorded by PageTriage extensions, or patrolling through other tools) to the number of articles created. It helps to measure how well a page review follows article creation.

Number of active patrols per day
This plot shows the number of patrol officers who perform at least one patrol action on a given day.

Percentage of patrol action by most active patrol officer, per day
This plot shows the proportion of all patrol actions performed by 25% of the most active patrol officers on a given day.

I hope this helps, let me know if you have any questions about them. Sincerely, Nettrom (talk) 15:24, August 29, 2017 (UTC)

Thank you Nettrom. This is much more useful even if the x and y axes can be done with better graduation. The graphics now all speak for themselves and clearly reflect the curves in the kiddy charts I produced last month that were not taken seriously. There is still no explanation of the massive decline in patrols/backlog hikes in mid-2016 that caused me to start putting pressure on running ACTRIAL, nor was there any idea about the strange coincidence of the sharp decline as my announcement to retire from my own role assumed a coord of all things PLTN. Kudpung ??????? (talk) 03:58, 30 August 2017 (UTC)

Kudpung Thanks again! I've added a monthly checkmark on the X-axis of the graph to make it easier to interpret, which adds a half-monthly mark between them. I choose not to do anything with the X-axis tick because the purpose of the graph is to expose the whole trend and allow you to predict what those values ​​are. If there is a need for a certain set of values, they can always be found in the TSV file I linked earlier.

From what I learned, the massive downturn on patrols in 2016 is because SwisterTwister is no longer on patrol (ref WMF reports from earlier this year), and the remaining NPPers do not compensate for it. As for why they are not on patrol again, you have to ask the patrols. Similarly, I do not know why there is a decline that coincides with your announcement. Cheers, Nettrom (talk) 14:33, 30 August 2017 (UTC)

Another update of the underlying data set generates updates from three original charts, and two of the graphs that display recent data with trends. I found a more subtle problem by counting patrol actions on some patrols that seem to have been counted twice. This has now been removed from the underlying data set and the graph is updated. The graphs affected by this are the proportion of patrol action for the articles, and the proportion of total patrol actions performed by 25% of the most active patrol officers. Please let me know if you have any questions. Cheers, Nettrom (talk) 17:33, 31 August 2017 (UTC)

src: upload.wikimedia.org

Deadline in 5 DAY

DannyH (WMF). There have been no convincing comments for a while. Can you confirm that WMF is on schedule for the 7th? thanks. (FYI: TonyBallioni). Kudpung ??????? (talk) 21:55, September 1, 2017 (UTC)

Hi Kudpung: Yes, security review for extension is completed today - you can see this ticket. I believe that's the only obstacle we need to go through. Morten ready, I'm ready. Kaldari is away now, but I hope he will confirm when he returns on Tuesday. We are on course for 7th September. - DannyH (WMF) (talk) 22:56, September 1, 2017 (UTC)

Thank you. Kudpung ??????? (talk) 23:04, September 1, 2017 (UTC)

@Kudpung and TonyBallioni: It looks like there's still no agreement whether to send newbies to a custom landing page or directly to the Articles wizard. We want to solve the problem before launching ACTRIAL, because it will cause a lot of chaos to change it in the middle of the trial. Should I turn that part into RfC for more opinions? Also, who will be responsible for the change to the last page of the Article Wizard? @MusicAnimal: Is that something you can do? Kaldari (talk) 17:26, 5 September 2017 (UTC)

@MusicAnimal: Oops, ping failed. Kaldari (talk) 17:28, 5 September 2017 (UTC)

Of course. I think we just need to add some user rights CSS classes around every option - MusicAnimal ^talk 17:38, September 5, 2017 (UTC)

This is quite interesting, as we are all on our way to see The OracleÃ, :) - fortuna velut luna 18:21, 5 September 2017 (UTC)

src: www.mariowiki.com

Countdown

I think the countdown clock posted above is misleading. It looks like we're ready to go on September 7, but we have not set a specific time for it yet. I do not want people to think that we promise to reach a certain moment. - DannyH (WMF) (talk) 18:31, 5 September 2017 (UTC)

Is the clock showing 00:00 UTC? Kudpung must wake up at 7AM Ã,;) - fortuna velut luna 18: 34, 5 September 2017 (UTC)

@DannyH (WMF) and Kaldari: The current time is set by Kaldari with these edits today. I think Kudpung took it from the checklist page, where ThereNime set it to 00:00. I have updated it on the page based on Kaldari time. TonyBallioni (talk) 18:36, 5 September 2017 (UTC)

Also, please note that it is not possible for us to give the right time. The 1 to 2 hour window as accurate as we can get since implementing new software is a complicated process. Kaldari (talk) 19:10, 5 September 2017 (UTC)

Kudpung will not even exist. He's a volunteer and has other things in RL to do. Kudpung ??????? (Talk) 22:57, 5 September 2017 (UTC)

Source of the article : Wikipedia