Topic on Talk:Page Curation

This is going to bite more newbies than the present system

6
WereSpielChequers (talkcontribs)

As currently designed the system doesn't allow you to fix a typo or add a category but leave an article unpatrolled. Instead it is designed to encourage the idea that patrolling is all about templating articles and tagging them for deletion. With the current system a large minority of patrollers have hotcat installed and a proportion of newpage actually get improved as part of the process, by designing out the most positive part of the newpage patrol process Zoom is likely to make the process even less welcoming to newbies than it currently is.

One theory for our decline as a community since 2007 is the shift from SoFixIt and crowd improving article to templating things for others to work on. By channeling patrollers away from hotcat, wikifying and typo fixing Zoom is liable to reinforce the trend that is undermining our community.

Ymblanter (talkcontribs)

I fully agree, but is not it what en:WP:NPP suggests to do right now? It says very clearly that if there are problems they should be templated, nobody suggests to fix them (even though many of the problems such as categories or references can be fixed very easily).

WereSpielChequers (talkcontribs)

We know from the survey that a lot of patrollers use hotcat, and there are things in the Newpage patrol instructions that go beyond just templating. For example "or add some appropriate categories to the bottom of the article. It is usually fairly easy to find two or three appropriate categories". I'd like to encourage more patrollers to fix things rather than tag them. Adding a category doesn't usually take much longer than tagging an article as uncategorised. So yes if Zoom cuts out the article improvement stuff and takes away the option of fixing a typo and moving on then I would be very concerned.

Ymblanter (talkcontribs)

This is a known problem, and I came across it many times when I was still working in ru.wp. Now I went to the end of the que of the articles which were not patrolled (on en.wp) and experienced it again. If I see a two-line stub which has no references, no categories, unclear notability, and no internal links, the easiest is just to leave there a bunch of templates and leave it patrolled (or not patrolled). As a responsible editor, I am doing more, for instance, I am looking for sources and insert them to the article. But in this way, it costs at least half an hour to improve (not even to extend, meaning adding new material) to a two-line article which is not immediately in my field of expertise. We can not really expect that all editors are so responsible that they start spending half an hour for each two-line article, and if we require this, they would rather not touch them, and the patrolling backlog would grow enormously (this is what happened on ru.wp and then some patrollers decided to set records and started to mass-patrol edits with an understandable drop in the patrol quality). Currently, I do not know how this could be solved.

WereSpielChequers (talkcontribs)

There are no easy answers to this one. But there are some partial answers:

  1. If you take off the deadline of it becoming automatically patrolled after 30 days then at least the pressure of working against a deadline falls away.
  2. Several of the templates we smear over articles should be automatically generated categories. Ideally hidden categories as our readers really don't need a category to tell them that an article has no categories or is an orphan. The templates that we show our readers really ought to be the ones such as unreferenced that are needed to warn our readers. If we gave our patrollers fewer templates that they could post and the patrolling screen had tools to easily de-orphan or categorise an article, then we might move a bit back to the SoFixIt mentality that built the pedia instead of the SoSmearItWithTemplates mentality that has so soured the project.
  3. One of our biggest problems with newpage patrol is that some patrollers think in terms of making sure that the badfaith stuff gets deleted, and others think in terms of whether articles are ready for mainspace. Both are of course worthy objectives, but you can't run both types of patrol with a binary patrolled/unpatrolled system. Or rather you can, but you get the current mess. The solution would be to move from a binary system to something slightly more complex that fits the patrollers mindsets. Instead of a binary choice give them an additional nuance. So the mark as patrolled box would become [Goodfaith/Ready for mainspace] if a patroller clicks Ready for mainspace then it gets marked as fully patrolled, if they click Goodfaith then it changes colour from yellow to green at specialnewpages and the next person to look at it just sees the option [Ready for mainspace]. At the front of the queue where the priority is screening out attack pages and other utter crap the aim would become making sure that everything was at least looked at and either tagged for deletion or marked as Goodfaith. I suspect many articles would get marked as ready for mainspace, but, and this is the crucial bit, everything would be looked at very soon after being created and I'd be surprised if the queue of completely unpatrolled "yellow" articles every ran to more than a few hours. The back of the queue of the screened "green" articles would continue as is, or possibly we might get more picky as to our minimum requirements for [ready for mainspace]. My suspicion is that deletionists and inclusionists alike would be happier under such a system - especially if we also had NoIndex set on unpatrolled articles. And of course the pressure on the "Green" articles could be changed from templating them for all sorts of improvements for others to do, to actually doing the necessary categorisation, wikifying and maybe even referencing before marking them as ready for mainspace.
Ymblanter (talkcontribs)

1. I may already start sounding like a notorious proponent of flagged revisions / pending changes, but for me this is what all this stuff is about. I was doing NPP for half a year in Russian Wikipedia (I was one of the first patrollers), then NPP were replaced by FR and I was doing FR for two and a half years, about 3K revisions per month (excluding autopatrol). For me, it is just a logical continuation of NPP, since (a) it fixes the 30 day problem - now every article can be patrolled, not just those which got into main space very recently - and this is logical, since we have a lot of old stuff which is inappropriate, and NPP/FR is a mechanism to coordinate effort of users looking at these articles systematically; (b) once one does this, one also realizes that even an article which is ready for the mainspace can be made inappropriate for the main space, for instance by being vandalized or by being subject to introducing copyrighted material or whatever. And then one would need to patrol it on a regular basis. And the most logical thing to do if one wants to coordinate this effort is to facilitate patrolling of a single edit - which is FR. Let us see how the RFC which was promised to us will run. As an intermediate solution, I do not have such a good knowledge of mediawiki, but I assume that unpatrolled pages remain unpatrolled forever, so that there should be a way of bot-creating list of pages which failed to be patrolled in 30 days and which still could be patrolled, but I do not know how to do it.

2. I agree with that. Is this smth which needs technical means to implement?

3. Actually, I think this ONE statement - that the article is almost ready for the main space - can be implemented by a template or by a hidden cat. But then it should be indeed understanding what NPP is. There was a similar situation indeed on Russian Wikipedia, when some users though a patrolled article is ready for the main space, and others marked everything which had categories and was not a blatant copyvio. The idea with three or more colours id excellent, but it obviously requires discussion and I do not know how easy it is to realize.