Talk:Task force/Community Health/Making Wikipedia a Happier Community

Dear Sjc, this essay is the best description of what's going wrong that I've seen so far. Policy-driven content: that's it. And it isn't restricted to the English Wikipedia. From my own experience (at the moment I'm having a break from wp-nl for similar reasons) I can add that similar blind enforcement of guidelines (as if they were rules) combined with deletionism is demotivating both newbies and veteran users at other projects alike. One has to have a skin as thick as an elephant's to be able to stand it for long. I'd recommend you to move this essay to meta, so that more people will read it.

Guidelines are okay, as long as they lead to the improvement of quality. Blind enforcement of guidelines destroys wikicommunities from within. Perhaps you'll find it interesting to hear that at wp-nl we have an unofficial guideline "bij twijfel niet inhalen" (literally: don't overtake when in doubt) which forbids neutral edits (edits that don't add or remove quality, see here), because they demotivate other users. Regards, Woodwalker 10:58, 28 November 2009 (UTC)

Thanks for indicating that it is also happening in the nl wiki, Woudloper. I speak enough Dutch to follow an argument, but not quite enough to really contribute there substantively these days. so I'll have a wander over there later to take a look. I will bide my time before moving this to meta if you don't mind, it needs to do a job here primarily and I certainly want to get a feel for what needs fixing in this before moving it to a wider area of readership. Sjc 15:57, 28 November 2009 (UTC)

Actually, the main problem I perceive at wp-nl is a bit different from that at wp-en. However wp-nl is (imho) a very nice case study to illustrate one particular risk for all projects. The Dutch Wikipedia is one of the very few projects that have enabled the system of 'patrolled revisions', a sort of (but not exactly similar) predecessor thing for the 'flagged revisions' system at wp-de (it was implemented 3 years ago, among the few other projects that have this system are wb-nl, wkt-nl, wkt-it). This has undoubtedly made patrolling and fighting vandalism at wp-nl easier. There is (in my experience) more oversight and more participation in vandalism patrolls at wp-nl because of the technical advantages patrolled revisions give. Any school-kid can help fighting vandalism, and as a matter of fact, many do. I know some examples of users that evolved, by careful guidance (IRC helps) and mentoring, from being notorious vandals into the most hard-working maintenance users at wp-nl (kids like to feel important/useful). This has led to the rise of a class of what I call 'maintenance users' (see my essay). They're good in one thing: fighting vandalism. When I looked yesterday the backlog of unchecked edits was just 3 edits and that isn't unusual. It's a very nice accomplishment that wp-de and wp-en can't equal. However, wp-nl is worse in fighting POV, imbalance and low verifiability. Battleground articles often stay disputed for years without any improvement.

Most of the maintenance users stay, become more and more respected and eventually become admins. Many quality users sooner or later run into a long-term conflict with one or more 'true believers' (see Piotrus' essay) and get disappointed because noone intervenes. At wp-nl, having made substantial additions to content is not a requirement for RfA (as at wp-en). Usually, the list of reverts of vandalism and list of articles 'checked' by the user (such a list is visible for any user in 'patrolled revisions') are studied and if they are okay, someone can become an admin. I have nothing against vandal fighting maintenance users (at the contrary), but in the current situation it's often impossible to have quality improving guidelines implemented because the majority of politically active users don't see the merits. The rating system (A, B, C, D) was deleted after a poll where a majority thought this makes TPs too long and confusing (at least one quality user I know left the project immediately afterwards). The guideline for verifiability isn't official yet because nobody seems interested in holding a vote. The same goes for some other important guidelines on what I called 'quality factors of content' in my essay.

The inner core/incrowd of the Dutch community now consists mostly of users doing 90% vandal fighting and other small maintenance, and very little users concerned with quality of content. New policy is at wp-nl, in your words, mostly policy-driven, not content-driven. Or perhaps even more precise: it is mostly maintenance-driven. This is demotivating for new users and users interested in adding content. The lesson we should all learn from this example is that to have content-driven policy, users that add quality content should get a fair share of the political power in a community. Woodwalker 07:06, 29 November 2009 (UTC)

I was wondering about the extent of the relatively neutral, trivial or otherwise recognisable maintenance edits I was finding in wp-nl WL prior to your comments, and their apparent timeliness; your notes above clarify matters greatly for me. This actually is, as you rightly note, a somewhat different trope from those I have so far mentioned in my essay, these being policy driven policy (in passing), policy driven content, content driven policy and content driven content. We now have a neutral and fairly unrewarding and unexciting trope to add to the mix, maintenance driven content. It doesn't, as a mode of production, add very much to the development of either quality, reach, range or extent of WP beyond keeping all things relatively quiet on the eastern front. nl:Wikipedia:Bij twijfel niet inhalen as an idiom does seem to sum up the modus operandi very nicely.

On a separate note, I am interested that you identify the need for users who add quality content to get a fair (or more appropriate level of) share of the political power. I can see why you say this, although I would be somewhat reluctant to envisage a situation in which we start to slice and dice users according to the merits or otherwise of their past edits. There are two main reasons for this: 1. A number of good edits in one article or area of expertise does not necessarily entail that all subsequent edits in different articles or areas of expertise are necessarily good (and by good I mean positive in that they bring content, quality or other value to the project). 2. Such a process is open to being played or downright abuse whereby an editor can build a measure of trust in an area and subsequently abuse that trust in another (as an extreme example, a user can build a measure of trust by making some useful edits to the subject of religion and then systematically use the built-up trust to POV blitz an article on or related to Darwinian theory). My take is that all edits should be taken on their own merits; a good edit should be recognised for what it is, a not-so-good or bad edit should be open to challenge and/or takedown or suitable emendation. This is the wiki way; it is equitable and democratic and if policy is non-intrusive and geared towards the improvement of content and quality, the wiki way works. The problem at present is that policy has now become so vast, complex and unwieldy that no single person can even begin to geta full measure of it. In fact, I would contend that in order to fully comprehend WP policy it would become a full-time job as rigorous and demanding as jurisprudence. We need to refine, simplify and make policy self-evident; where the policy is not clear, well-documented and self-evident, we should simply fix or excise it.

As an example of this, see [1], discussion pertaining to the centrepiece of this essay. The policy is self-evident, if you actually read and interpret it correctly. The non-notability policy is not well-documented, and has ambiguities and requires more than three spare brain cells to understand the implications of the qualifying rubric. The corollary of this is that policy fetishists see this as ammunition in their attritional war against anything they take a dislike to and abuse policy, process and commonsense in their headlong rush to enforce policy, often, and as in this case, incorrectly. Sjc 08:39, 29 November 2009 (UTC)

You're right that some people see it as ammunition. Funny enough, authonomy would meet the guideline just fine. The problem is that someone can just use the guideline to proclaim anything non-notable without going to any effort. It puts too much stock in the !voting process (which can be disrupted by one faction or another), and not enough stock in whether the article is actually fixable. Randomran 21:06, 29 November 2009 (UTC)

(@Sjc): Nice to hear that people outside the local community are interested in the local language projects (Dutch has some 25 to 30 million speakers, most living in the same small area in western Europe). Actually, I don't think cases like authonomy are the worst problem. Wikilawyering (bending guidelines to one's interpretation) is probably unstoppable and perhaps we will at some stage all be guilty of it (knowingly or unknowingly - in this case unknowingly when we AGF in Orange Mike). Fortunately, wikilawyering is often easy to unmask. I acknowledge the discouraging effect it can have on good contributors though.

Something more about this 'maintenance-driven policy'. As I wrote in my essay, maintenance users are okay, since they increase the overall quality of a project. However, if we take a look at the list of quality factors I made, we see that they focus on the 3rd and 4th groups: requirements of form and requirements of the project. Examples are improvement of categorization, spelling, template stickering, etc. This will help the reader, yet it doesn't create more content. Growth of content is accomplished mainly by increasing the first and second groups (requirements/factors of content and demand). To keep growing, a project needs quality users. Now back to the Dutch example. My worst fear is that maintenance-driven policy creates a negative feedback loop, when it has the effect of only creating more maintenance users. I fear this is happening at wp-nl. If this is true we will end up with a project where only minor maintenance edits are made and growth of content has stopped. The same can hapen with policy-driven policy: it encourages politically interested users to become more active in making yet more policy, while it discourages other types of users.

Compare maintenance-driven policy with policy-driven policy, and you will find some differences. The long-term effects of both models are the same though: creativeness and growth of content are stopped and quality users are discouraged. I use the word "models" since they are probably not perfectly valid for any single project, yet most larger projects will have some characteristics of both. Woodwalker 07:49, 30 November 2009 (UTC)

Surprised

I am a little surprised by the assertion that the case of authonomy is an example of something that is exceptionally bad or something that is a recent development. I have seen and been subject to much worse (including insertion of extreme PoV's and routine violations of NOR) and much more systematic abuse, and that started over three years ago. In many ways Wikipedia is a huge machine, and it is not really aimed at producing usable (verifiable) content on notalbe topics. Wikipedia does contain some very good content, but only in places where users devoted to good contents succeed in making the difference. In many areas the machine is just producing dust. - Brya 08:12, 7 December 2009 (UTC) (I agree with Woodwalker that the Dutch Wikipedia tends towards a machine run by "maintenance users", and that this machine has weird dynamics. Not so sure if this is really different from the English Wikipedia, except that in the English Wikipedia other groups have a stronger representation. The "bij twijfel niet inhalen" is a translation from a part of the English Wikipedia Manual of Style; it did not originate at nl. - Brya 08:12, 7 December 2009 (UTC)

@Brya: Thanks for pointing out its origin, I wasn't aware of it. Woodwalker 08:15, 11 December 2009 (UTC)

The point about authonomy is that it not only presented to me the extreme abusiveness which is now present in en.WP but the systematic and coordinated efficiency of the onslaught. To be speedied within 1 minute is almost a mindless response. It then got another speedy in an hour or so. Then the tag fest. It was certainly enough to tell me that I was well out of there. You're right in identifying the problem appearing about 3 years ago, but I think I, and I suspect many others of fairly benevolent intent, have shut our eyes to the problems which exist for a very long time. I've worked on the project since pretty much its outset but I rather feel that I have been driven out by the deliberate policy creep, abuse and misuse of policy and insidious wall of hocus=pocus bureaucracy which are a million light years from the original philosophy and raison d'etre of the project. Sjc 18:53, 9 December 2009 (UTC)

Mindlessness and Wikipedia have been fast companions for as long as I have known Wikipedia. Terrible things have happened always. The question rather is one of frequency. My guess would be that it is decreasing, as less editors make for less misbehaviour. - Brya 16:36, 10 December 2009 (UTC)

I think this thread leads us to the same parent problem that was identified during the discussions of the quality task force. It's the problem of having the wrong brand. I like to quote Bhneihouse:

Because the brand statement is not “encyclopedia” but rather “encyclopedia built and run by the community” there is no way that anyone can be excluded and have it be fair and equitable. If the brand was simply “encyclopedia” then it would be fair to exclude those whose contributions interfered with the mission of “encyclopedia.”

It is the problem of not having the goal of building an encyclopaedia made clear enough. Woodwalker 08:15, 11 December 2009 (UTC)