Wikipedia:Strategic issues with core policies, guidelines and structures

Unless indicated otherwise, "Wikipedia" refers to the English Wikipedia.

Wikipedia is a huge success story but it also has significant problems. Of course, to some extent, "problem" is in the eye of the beholder. This attempts to rise above that and identify problems that harm or threaten Wikipedia's mission as an encyclopedia. Of course some are narrower and more specific problems that are also the causes of other problems. A weakness in a specific area is both a cause of a problem and a way to fix it. In some cases a solution is closely paired with and blended with a statement of the problem and is combined as such. In other cases, a solution may help solve many problems and so is listed separately.

Executive summary: 5 changes that will 80% fix Wikipedia's biggest problems

1. Create a new type document that defines what Wikipedia is and how it works. Ultra-consensused and authoritative and general....being such it is not to be invoked prescriptively but instead guides and influences policies and guidelines and other items. The first of these will be the current 5 pillars. Somewhere in one of them we'll define editorial discussion and discretion as central processes as central to what we do. Trying to ignore that and instead be guided only be game-able rules is a source of problems. An en-wikipedia constitution might be one of them.
2. Modify WP:Verifiability to incorporate actual reliability into the RS criteria. This criteria would be expertise and objectivity with with respect to the item which cited it And acknowledge that there are degrees of strength of sourcing. Say that the more that the veracity of the item is contested, the stronger the sourcing must be, and vice versa. Add that verifiability is a requirement for inclusion, not a reason for inclusion.
3. Modify definitions of WP:Reliable source to include with respect to the item which cited it And to acknowledge that there are degrees of strength of sourcing and that anything that states otherwise is incorrect.
4. Completely re-write wp:weight. A few paragraphs will do. It is an immensely crucial part of our policies, yet it is fundamentally unusable, very easily game-able, and badly outdated and works against the objectives of wp:NPOV
5. x

Main Problems

Wikipedia is a mean and vicious place for editors

Wikipedia is a mean and vicious place for editors. The worst of it is not from violating the letter policies and guidelines, it is from weaponizing policies, guidelines and other Wikipedia structures. And the second worst cause is other mis-uses of these items. At best it is rough-and-tumble and mean place for a majority of editors. For newer editors who have not yet learned the 1,000,000 plus words of the official guidelines and policies, they will be consistently "violating" those policies and guidelines plus a bunch of other unwritten ones, and will be told that they are. And if they operate in any contentious areas, they will be attacked by vicious warlords and their posses/militias who have learned how to weaponize policies and guidelines to conduct their warfare. Causes/solutions:

• Create recognition of this. Since it can be easily disguised as "just following / enforcing Wikipedia's rules/principles" so painting a Scarlett Letter on the practice is needed

Systemic bias exists to the point of impairing informative coverage

Bias on US political and US culture war topics is a problem not simply because it is bias, but because it has reached and extent to impair providing informative coverage. To address this, it first must be recognized that there is so much subjectivity and complexity in defining any high "unbiased" standard that such does not exist. Nevertherless, meeting a much lower bar is achievable and needed. The lower "bar" is of not having one "side" of that battle influence Wikipedia's content so much that it degrades informative coverage or that it garners such widespread disdain in the outside world that it damages Wikipedia reputation as a source for enclyclopedic information, particularly in those areas. Currently Wikipedia often fails at achieving that lower bar.

The factors leading to the perceived left bias also degrade informative coverage of topics

The factors leading to the perceived left bias also degrade informative coverage of topics

While the WP:Notability-system criteria for existence of an article 90% serves Wikipedia's mission well, but 10% is a high failure rate

While the WP:Notability-system criteria for existence of an article 90% serves Wikipedia's mission well, but 10% is a high failure rate. The WP:Notability-system is a complex kludge. Since understanding it is a necessary condition to improving it, such is a barrier to improving it. Once that hurdle is passed, the complexity itself is will make it difficult to improve.

The admin system is broken in several ways

The admin system is broken in several ways. There is a lack of new admins, and also a lack of admins suitable to handle complex, high-expertise and difficult situations.

With the RFA process broken down, right now the main criteria for who is in the admin cadre is "got in back when it was easy". And the second criteria for new ones is "kept their head low" / having avoided contentious situations. Impacts have already been felt and will get worse.

The root of the problem is that the admin function conflates these two things:

• Type #1"no big deal" tool belt functions with
• Type #2 other "big deal" "judge" type immense powers given to these folks (such as being able to sanction established individual editors, close complex and contentious discussions), defaco immunity from review of 99% of their behavior, and immunity from review of bad decisions that they make. This is conferred also by policies and practices, not just by software definitions of the tool belt.

With #2 blended in it really is a very big deal and rightfully a very tough process.

Step 1 is to separate them. This can easily be started informally. Start a "Yoda" (my word for it) class of admins with extremely strong and thoroughly proven credentials of thoroughness, fairness, avoiding emotional responses or decisoin making, self-control, excellent decision-making, immense experience, immense knowledge of guidelines, policies and how Wikipedia actually works. Also a record or never mis-using their admin clout in a pissing war. We only need about 20 or 30 of these. Evolve to a situation where they handle the tough behavioral cases and they become a routine appeal point for bad decisions by admins. This will solve a wide range of problems and make it closer to "no big deal" and 3/4 solve the RFA problem.

The other 1/4 of the RFA problem can be solved by structuring 90% of the discussion. Determine what qualities are needed, and direct 90% of the discussion to whether they have those qualities.

There's a big gap at the level just below arbcom

As a venue for handling problems, Arbcom is the highest level in most respects. This position derives from many attributes. "Just below" arbcom there is a huge gap. The next level "down" is where any person with the tool belt can make the decision.

Undisclosed paid editing is a big problem with many impacts

Undisclosed paid editing is a big problem with many impacts.

Wiklawyering is far too easy to do and far too effective in Wikipedia

Wiklawyering is far too easy to do and far too effective in Wikipedia. Misuse of policies and guidelines is more common than violation of them. As a result it is widespread

Even when things are going well WWF does some harm and prevents proper allocation of resources

Even when things are going well WWF does some harm and prevents proper allocation of resources.

Structural problems with top level governance and WMF are an existential threat to Wkipedia

Structural problems with top level governance and WMF are an existential threat to Wkipedia. These weakness mean that it's possible for things to go things to terribly and unsolve-ably wrong.

WP:NPOV often fails at or works against its objective and needs improvement to better define & achieve its goals

WP:NPOV often fails at it's objective and needs improvement to better define & achieve its goals

It is very hard to learn Wikipedia

The rule book that runs Wikipedia

The rulebook that runs Wikipedia consists of approximately 73 policies and approximately 280 official guidelines for a total of approximately 353 sets of official rules. Each of these sets of rules ranges in sizes from a few words up to approximately 25,000 words amounting to approximately 1.4 million words of official rules. But even that is only a start. If one were to read and learn all 1.4 million words of policies and guidelines, one would still be less than 1/2 way to an understanding of the actual rules that Wikipedia operates by. This is because:

1. About 1.4 million words of crowd-sourced rules in 350 documents are inevitably going to have a large amount of overlap, conflict and wide variations in potential interpretations.
2. There are even thousands more documents called essays and "supplemental pages" and a few of those are also often treated as rules. There is no list of which of these thousands are, so you would need to learn that from experience.
3. Individual projects (which are groups of people specializing in individual topic areas) write their own guidelines which they consider to be applicable. They will often consider these to be rules that everybody else outside of their group is supposed to follow, and newer editors are often misled to believe that such is the case.
4. There are some unwritten accepted practices and interpretations which are only learned through a large amount of experience.
5. Most importantly, rules are often stated as an overly broad statement, but then mitigated / reduced with softening qualifiers (e.g. "generally", "usually")

What makes this seemingly impossible situation work most of the time is the fuzziness of the Wikipedia system. This includes doing things by consensus, weighing multiple considerations, giving more obscure rules less weight, deciding to follow rules only in the situations that they are intended for (technically, using their softening qualifiers to not invoke them in other situations.)

While this works most of the time, this leaves the situation wide open to abuse. This is amplified by the fact that in Wikipedia, every editor is empowered as a policeman. Imagine a speed limit of 30 mph. The intent and practice is that tickets will be written for 5mph or more over the limit, not 1 mph over the limit. But with every editor empowered to be a policeman, the can harass an editor or knock material out of an article to tilt it by saying that 31 mph is a violation, and then get immunity by saying "just following the rules, if you don't want trouble, just don't break the rules".

Even though this system works most of the time, the problematic situations are significant and need repair.

Mis-use of core policies is a larger problem than violation of them

Violation of policies and guidelines is easily resolved and generally is. Misuse of them by wiki-savvy persons is difficult to resolve and so is often not resolved. So, at any given moment, the latter are much more prevalent. When it serves their purposes, and they can get away with it, people will commonly mis-summarize and mis-use policies and guidelines. This is not an unsolvable problem (e.g., "you can't stop people from ...."). Flaws in how these are written is the cause of the problem, and improvements in how policies and guidelines are written will substantially improve this problem.

Analysis: the mechanics of content policies and guidelines in action

Policies and guidelines operate at two main levels:

1. The highest plane, where they express the general intent of what should be, and people who seek guidance follow that.
2. The nuts-and-bolts operative plane. This is what takes over when there is a dispute....operative mechanics and wikilawyering prevails over the general intent.

Lets look at #2: Everything is in Wikipedia articles, and every action and change in a Wikipedia is the result of one of these three things:

1. Addition of material
2. Deletion of material
3. Modification of material (which is simply a combination of #1 and #2)

So, the only way that content policies and guidelines affect content is by affecting these three actions. Most operative clauses of core content policies and guidelines operate only by placing conditions on the presence of content. (one rare exception is the balancing provisions of wp:npov) So, to simplify even further, content policies and guidelines act by influencing against the addition of material or for the removal of material. Here are 3 different levels of strength of this "influencing":

1. Give carte blanch for removal/exclusion. I.E. empower an editor to remove it, even over the objections of others
2. Give a strong argument for removal/exclusion. Not #1, but gives an argument that would weigh heavily in a discussion
3. Give a not-strong argument for removal/exclusion

So every piece of content must comply with dozens of policy/guideline conditions. So, meeting any one of those (such as the sourcability/sourcing requirement) is not a force or mandate for inclusion, it simply fulfills one of the many requirements for inclusion.

Solution Wp:VER / WP:NOR

WP:ver and wp:nor are discussed in tandem here because structurally they are mostly duplicates of each other.

WP:ver and wp:nor need additional source metrics, and a way to apply them

Preface on terms etc.

This considers the current definition of "RS" to be per all and only source criteria defined in wp:ver and wp:nor. So it does include primary/secondary/tertiary related suitability as defined there, but does not include non-policy reliability criteria such as actual reliability or that reliability is strongly context-specific (= varies with context) The latter are often invoked or discussed at the RS noticeboard, but are largely missing from policies and guidelines, and are erroneously ignored and countermanded by overgeneralized listing of sources as being unreliable or reliable.

Main discussion

The current structure of wp:ver and wp:nor essentially provide two criteria for sources. One is the "reliable source" criteria the core of which that it generally requires an internal layer of review, with "publishing" often considered to be this. The other is primary/secondary/tertiary, with restrictions on the use of secondary and tertiary sources. With respect to the wording of wp:ver and wp:nor, fulfillment of both of these criteria is considered necessary and sufficient for a source to be considered suitable for the use. This has significant problems. "RS's" are often unreliable, and sources with very high real world reliability often are not classified as wp:"RS's". Further, the clear intent for requirement of RS's in wp:ver/wp:nor is suitability to support material which cited it. Pretending that "reliability" is independent of the topic / context is not realistic.

Solution: enhance the definition of reliable sources and acknowledge variable strengths

Add the following two criteria to all definitions of wp:reliable source: Objectivity and expertise regarding the item which cited it.

Instead of making the each of the (now to be 4) criteria categorical, acknowledge that:

• The sum of the four criteria with respect to item which cited the source would be considered the "strength" of the citation.

• The more controversial and challenged the material (which cited it) is, the greater the required strength of the citation, and vica versa.

Here's some wording that takes a small step in that direction:

• Controversial claims require stronger sourcing. Two additional measures of this are the objectivity and expertise of the source with respect to the material which cited it.

The tiniest change with the largest benefit

The following tiny change would have substantial benefits in all of these areas:

• Reduce abuses of wp:ver & wp:nor while not affecting their strength even 1%
• Substantially reduce warring at articles
• Reduce wiki-lawyering
• Increase effectiveness of wp:npov

The change

Add the following somewhere in wp:ver & wp:nor: When tagging or removing material for being uncited or insufficiently sourced, please indicate that you have a concern or question aside from citing and sourcing.

WP:nor and wp:ver should be brought in line with how successful articles are created

Most of Wikipedia is built upon "violations" of wp:nor, if it is taken literally. Nearly anything except a direct quote violates a literal reading of wp:nor. (e.g., summarizing is a violation of wp:nor-taken-literally.) While most do not recognize or admit it, the way that most of Wikipedia is actually written is from consensused knowledge (knowledge = an integration of hundreds of sources, with sources at hand in mind) and THEN sourced per wp:ver & wp:nor. Most of the time this disparity between the rules and reality works fine. But on contentious articles, any wiki-lawyer warrior can roll out the oft-"violated" rules to exclude the other side's content. Using this they can also shut down the above process which works elsewhere in Wikipedia, and there is no mechanism to get that process back in place because it is technically "illegal". That is a big part of why most or all contentious articles in Wikipedia are eternal messes and eternally unstable. It is also a substantial contributor to the problem that experts don't edit in their field in Wikipedia.

This process also includes collectively striving for accuracy. While accuracy is implicit in the highest level statements (by-laws, mission statements) it is missing from and not acknowledged in policies and guidelines.

Solutions

Align the policies and guidelines with reality in this area. This includes the mild level or synthesis / summarization that Wikipedia is built on. But at the same time, do this without lowering the bar on admission of questionable material This seemingly complex task would actually be VERY easy to accomplish. A challenge of the material is still sufficient grounds to require sourcing per the current rules and definitions. The change would be that such a challenge should include a question on the challenged beyond just saying that it is unsourced. The challenger needs only raise the question/concern to invoke the sourcing requirement, they do not have to discuss or support the question/concern further in order to the cite request to stand

Mis-use of wP:ver to say that verifiability is a reason for inclusion

Though it is not in any policy or guideline, the fact that material is wp:verifiable is often used as an argument for inclusion. The policy weakly and mistakenly tries to mitigate this by discussing inclusoin/exclusion of verified material. A weak 1 sentence attempt at a giant topic. Instead it should say: Verifiability is a requirement for inclusion, not a reason for inclusion.

WP:NPOV

Wikipedia articles on contentious topics generally remain in a permanent state of conflict and instability. WP:npov is not currently sufficient to significantly support resolution of these. Certain fixes would require changes in other policies, especially source and sourcing criteria. But others can be helped by improvements in wp:npov:

Wp:npov needs text pointing out the most frequent forms of POV'ing

The most frequent forms of POV'ing slips "under the radar" of wp:npov, i.e., it gives little direction or guidance for those areas. Partial solution: Include something like the below in NPOV. Merely recognizing, understanding and spotlighting these other forms of pov'ing will help the situation.

A common assumption is that most POV questions on articles arise where there are conflicting viewpoints on a particular question, or point made in the article. This is mistaken; a minority of pov issues are of this type. Here are three more types, with an example of each:

• Selection of particular angles to cover. It matters not that the coverage of that particular angle is objective or balanced, the selection itself tilts the article.

  Example, if there is a move to cut government programs and taxes, choosing to objectively cover a case where someone would be hurt by a program cut tilts the article one way, choosing to objectively cover a case of a person hurt by high taxes tilts the article the other way. So it matters not that the coverage of the particular angle is balanced; the choice of the angle to cover creates the imbalance.

• Quantity of content The quantity of content on positive vs negative topics regarding an organization tilts the article one way or the other. Note that this is true even if the coverage of those chosen topics is unbiased.
• Insertion (association by mere presence in the same paragraph or article) of powerful material

  Example: "Priest John Smith said that he is a homosexual. Smith critics noted that the XYZ study concluded that allegations of homosexual child molestation by priests is 9 times more common than allegations of heterosexual child by priests."

  Even though the second sentence provides no info about or even any accusation of John Smith, the presence of child molestation information in the John Smith article implants the association in the mind of the reader. Currently the only defense against this a weak one under wp:nor, basically saying that the juxtaposition of the two items is synthesis, or the weaker one of trying to use the toothless wp:undue.

One of the two biggest holes in wp:npov

The biggest hole in wp:npov is that it basically only addresses cases where the material espouses opposing viewpoints on a particular topic. Material which is put in for effect but which does not espouse a viewpoint slips under the wp:npov radar. The only section that slightly addresses the latter is wp:undue, and, as written, it is ineffective in this area.

The other of the two biggest holes in wp:npov

This that wp:undue is basically toothless in disputes because its main guideline for implementation (preponderance in sources) is really not practically usable. It has also become outdated because most media have become biased in certain areas such as current politics. One solution would be to incorporate other metrics into the guideline. The objectivity and knowledgeably of the sources with respect to the topic should be added to the raw preponderance criteria. Also, include directness-of-relevance as a criteria to be taken into consideration.

The current policy works at the "general principle" level (as a general guiding principle) but when you get to the "operative mechanics" level (which is what's used when there is a dispute) it is absolutely unusable. So a usable form of wp:weight is probably the most important missing policy in Wikipedia. I don't have an answer, but a good starting point would be to list things that should be taken into consideration in the "applying wp:weight" equation. Most of these are sprinkled through the above conversation, but they might include:

• Meeting the "floor" of wp:rs criteria (this is the current criteria)
• Impartiality of the source with respect to the topic at hand
• Expertise of the source with respect to the topic at hand
• Prominence of coverage
• Degree of focus of the coverage on the topic at hand. (e.g. source just mentions the topic vs,. being about the topic)
• Prominence & credentials of the source
• Degree of distribution of the source. (e.g. circulation)

Problem and solution: Degree of relevancy isn't included in NPOV guidance, and needs to be

Wp:undue does not provide sufficient guidance to resolve contentious articles and generally fails on these. Including degree of relevancy as a factor to consider in those situations would help. This can be accomplished by upgrading WP:Relevance to a guideline. In also includes a framework to determine and name the degree of relevance.

Exploration of situations that would benefit from this

WP:npov seems best designed only for the classic POV case, where there is a statement which purports to be objective fact in dispute. But the far more common case is where POV warriors seek to leave an impression on the reader via the quantity and nature of content which leaves the desired impression. This may be:

• On the topic of the article, or
• Where the POV promoter inserts material into the article to further a POV on a different topic.

An example of the "on the topic" type would be if Rush Limbaugh announced that Barack Obama is the worst president in the last 100 years, and many newspapers reported (simply) that he made this announcement. And then an editor puts a section on this into the general Barack Obama article. Technically, the editor is not inserting/citing/having to argue the "worst president in 100 years" statement, they are just saying that Limbaugh said this. They just want the very real impact and impression of the presence of "worst president" type words in the article. A second example is that if John Smith, a person who is a second cousin of Obama is convicted of child molesting, and the conviction is covered by several newspapers in a matter-of-fact manner. And an editor places a section into the general Obama article regarding that topic. They make no other argument that needs defending, they just want the impact of child molestation related material in the Obama article and its juxtaposition with Obama material. Most would say that these should not be in the article. And, if there were a large amount of such material in the article, most would (intuitively) say that such POV's the article. But policies and guidelines provide little guidance regarding this. The sourcing is not only on wp:solid ground, the coverage really can't be questioned, as it was matter-of-fact regarding these matters. Ditto for the "objectivity" of the text put into the article, it is simply matter of fact coverage of Limbaugh's statement and the 2nd cousin's conviction.

Probably the policy/guideline most looked at for guidance on this would be wp:undue. But it is oriented towards covering opposing views on a particular statement. In these cases, the "statements" are just what was said in the speech, and the facts of the conviction. It gives guidance only on coverages of two sides of an issue. But there is no debated "issue" in this material, as it is a statement of facts regarding what Limbaugh said and of the conviction and of the relation of the child molester to Obama. Even worse, wp:npov says what can be interpreted as "must include" for these statements.

Solution

Work is at the essay wp:Relevance. Upgrade that essay into a guideline.

Wp:npov needs to include more guidance on section titles

Section titles tilt this inclusion of information in an article. They influence the article to include a greater amount of material defined by the title. Example: John Smith kicked a dog once, a long time ago, and also runs an animal shelter. In the John Smith article an editor creates a "Controversies" section. This tilts the article towards inclusion of a greater amount of negative material on John Smith. It might tend to give a section on the dog-kicking incident legitimacy for inclusion which it might not have otherwise had. And it could be used to prevent another editor from including the dog sheltering material to provide balance on the topic of Smith's treatment of animals. The removing editor can say that the dog sheltering material was removed because it is "not a controversy".

Solution

For contentious situations, section headings may be created only for material that could pass the wp:undue test for inclusion in the article without the section heading. Further, only material that can remain in the article without having it's suitability "propped up" by a section heading remains. Basically, this means that material must "stand on its own two feet" regarding justification, without such being "propped up" by the section title.

Analysis of compound criteria titles and articles

There is variability and ambiguity on what the confluence of two or more criteria in an article title means. This variability / ambiguity affects several areas including:

• Evaluating whether or not the subject meets wp:notability requirements
• Evaluating whether or not the article complies with other policies and guidelines
• Evaluating what is / isn't or should / shouldn't be within the scope of the article

This doesn't provide the answers, just an analysis and framework to assist in that process. Here's a breakdown of the purposes of compounding criteria in these articles.

1. Where it is clearly only to narrow the scope to a suitable size. For example: North American traditional folk music. These follow (only) the common naturally assumed narrowing methods such as by a country or time period. For example, Folk music performed by bands with three members is not such a "only to narrow" situation.
2. Where the title/article is about a possible, presumed or actual cause-effect relationship e.g Major plane crashes in unregulated airlines
3. Where the selectivity of the title/article is significantly informative or of interest to readers e.g. Albums that sold over 5 million copies
4. Where the confluence is simply interesting without a significant other purpose e.g.Beautiful actresses who were ugly when they were children
5. Where the compounding does not significantly serve any of the above purposes e.g. Baseball players who hit a home run during the third inning of their first professional game
6. Where the title/article primarily serves a purpose of making something or someone look good or bad e.g Child molesters who voted for John Smith

For situation #2, the wording of the title follows common practice and good prose, but taken literally conflicts with the purpose of the article. A precise title would be the long awkward : The possible cause-effect relationship between an airline being unregulated and having large scale plane crashes. Instead, this is implied in the title and then covered in the article. But this presents an ambiguity. Is the article also coverage of the plane crashes themselves? Whether or not the plane crashes have their own articles can be relevant to this question. The answer can be "no" or "yes, list only" or "yes, brief coverage" or "yes, full coverage, even if there is overlap with another article." For #3 or #4 similar questions arise. Of course #5 & #6 should not be Wikipedia articles but we need to recognize them in order to avoid them.

Wikipedia is a very mean place

Despite immense efforts to be the opposite, Wikipedia is a very mean place. Most of the efforts work at one level (brazen problems mostly by newbies) and backfire at the level of people who have acquired the wiki-savvyness to use the system to conduct warfare. For example, they have learned how to use the wp:civility guideline to conduct vicious warfare against someone.

WP:civility

The policy that is considered to be the one that generally deals with nasty behavior towards other editors is wp:civility. But it is ineffective towards most of the viciousness and nastiness in Wikipedia. The weaknesses fall into 2 general categories:

• The real world meaning of being "civil" is a standard of behavior that is unnecessarily and unrealistically high in a very narrow field (generally, gentleness of a conversation) and which
• ignores a wide range of nasty behaviors

A quick metaphor to illustrate: Civility says that you can't use rough words on someone, but it doesn't say that you can't put a knife in between their ribs. Let's delineate some of the "missing coverage" areas.

Mis-restating something that the person just said

Restating what someone just said should be of your sincere "take" of what they intended to say to see if you understood it properly. Unfortunately it is very often used to MIS-state what they just said in a way designed to make them look stupid, fanatical, in violation of policies or norms, or otherwise to make them look bad:

John says "The XYZ policy applies to article space, not talk pages, so what I did was OK"

Joe says: "So you are saying that your personal opinion overrides policy? That's not how Wikipedia works"

John says: "Those two editors appear to be exerting wp:ownership over the article"

Joe says: "So you are claiming that there is a conspiracy to control the article?"

Untangling the COI Gordion knot

What was previously the main definition is a universal masterpiece, even if it does not define or handle the nuts and bolts: "When advancing outside interests is more important to an editor than advancing the aims of Wikipedia, that editor stands in a conflict of interest" This is the definition used here; other conflicting definitions exist even within the current wp:coi guideline.

2. The COI equation

• 2.1 On one side of the balance is the strength and type of COI-type influences present for an editor on a subject
• 2.2 On the other side of the balance is the strength of personal character etc. that could self-restrain themselves against COI editing. Also some Wikipedia mechanism which discourage COI editing.

Structurally the operative parts of wp:coi are a blend of these three things:

• 4. Attempts to define "presence of interest" situations which are an issue or risk. This has several components/measures

• 4.1 Degree that certain benefits are concentrated on the editor:

• 4.1.1 High Examples: Payments to the editor, self promotion of the editor (e.g. their book), the editor receiving a promotion at work.
• 4.1.2 Medium Examples: A regular employee of a large company, helping the company do financially better,
• 4.1.3 Low Example: A person who has children benefiting from promoting a child welfare cause.

• 4.2 Extent that benefits to the editor are monetary vs non-monetary. High = e.g. pay for editing, increased sales of the editor's book, a promotion for the professor covered in the article. A non-monetary benefit might be the pleasure of furthering one's (outside-of-Wikipedia) cause.
• 4.3 Varying degrees of knowledge by Wikipedia of the factors relevant in the COI-editing equation.
• 4.4 Does the situation involve control of the editor by another party? And if so, does the controller have a high-risk-of-COI relevant to that control. A professor giving a grant or good grades to a student to write an article on the professor has a hich-COI risk. A professor giving a grant or good grades to a student to do good writing on unrelated articles does not.

So, there an be different linked "strings" of answers that go through #4,#5,#6. E.G. "If ABC is present, then XXX applies" "If XYZ is present, then ZZZ applies"

• 7. Discussions have centered on whether particular situations are problematic on their own merits. Are there certain behaviors that we want to discourage for strategic purposes, even if they are not necessarily individually problematic on their own merits? For example, Wikipedia could say that it has a strategic interest in avoiding a paid editing cottage industry from growing, and that it wants to flatly ban all paid editing, (thus forcing such to be secretive, which vastly reduces it) even if there is no actual COI and the editor behavior is otherwise perfectly fine.

Framework notes

• Decisions should keep in mind that self-declaring a potential-COI situation often, for legitimate reasons, not occur. So the plan should not heavily depend on that.
• Being willing and able to set influences aside when editing is a quality / situation that some people have and some don't, and with varying degrees. For example, when a person paying them is guiding their Wikipedia work. So some people may WANT guidelines or firm rules that apply to themselves.

After deep analysis, it boils down to being simpler than it looks

I've been analyzing this for a year and came out with, once the haze is cleared away, it being simpler than it looks:

1. If interests (which conflict with the goals of Wikipedia) are influential enough to overwhelm the person's fortitude then, if they edit unconstrained, they will edit against the golden rule at the top of wp:coi.
2. If #1 is the case, and they disclose that they have a high-risk-of-COI-editing situation at an article, their reduced clout combined with the normal editor and article processes will prevent problematic editing. It will go smoother if they follow the wp:coi editing procedures, but they are not essential to this. However, I've learned that some COI editors WANT the guidance of COI-editing rules either because they simply want the guidance and a rulebook to follow, or to deflect improper requirements and expectations form the persons paying/controlling them.
3. If #1 is the case, and they don't disclose it, then they are beyond the reach of policies and guidelines (and so such are worthless in that case) and problematic COI editing will occur.

So, what matters above all is encouraging disclosure in every possible way if they have a high-risk-of-COI-editing situation. Including letting them simply say "I have a high-risk-of-COI on this article" with no further questions (which could tend to out them) on that then allowed. And the top of that "high-risk-of-COI-editing" list is paid advocacy. So a more narrowly written version of the proposal of this article, but which requires disclosure (not ceasing of editing) should also be included.

COI Definitions

We need to acknowledge that "COI" is given two completely different conflicting meanings within Wikipedia, (even within wp:coi) and that neither of those matches the (third) real world meaning. The two Wikipedia meanings are:

1. A situation out of compliance with the "golden rule" definition exists. I.E. when the strength of conflicting-with-Wikipedia intersest is stronger than the person's fortitude
2. The trappings of high-risk-of-coi situation exists. Paid editing that serves the interests of the payer is an example of this. wp:coi also hints at a ridiculously wide range of other situations as being a COI, while completely missing other huge ones.

The third definition is the real world one: COI generally means a financial interest that a person has which will be directly impacted by a decision that they are involved in making, where, in that role, they are supposed to be acting on behalf of an organization or constituency. So it is "trappings-based" but very narrow.

Good faith challenge

Every Wikipedia policy or guideline has an intended purpose or intended purposes. Sometimes policies and guidelines are used for reasons which are contrary to or outside of their intended purposes. Examples of this are to POV an article, conduct warfare, or indulge bad tendencies. Some of the latter instances are called wikilawyering. Adding the following would have an immense positive impact: When raising a policy or guideline-based issue about content in an article or about an editor or their behavior, please include a Good Faith Concern or Good Faith Challenge that relates to a purpose or intent of the policy or guideline or other concern outside of the wording of the policy itself. This will help reassure other editors that you are not wikilawyering and strengthens your case for that action.

Such expressions may for example, include expression of concern regarding the accuracy, trustworthiness or neutrality of material, or the impacts of editor behavior on others. This is just an initial expression that provides extra assurance; the person does not need to have or win a debate on their concern in order for their statement about a policy or guideline to have validity. So if someone says "I deleted this material because it does not comply with policy XYZ, and I think that it may be inaccurate", debating the "I think that it may be inaccurate" is not required and must not displace application of the policy as the crux of the conversation.

No abuse

Processes (such as Arbcom, wp:an, wp:ani, RFCU) should not unfairly abuse editors, and should not be used to conduct warfare. Trying to cleverly, unfairly use these venues (in combination with policies and guidelines) to "get" or deprecate someone is a common and destructive problem in Wikipedia. Unless they involve items that must necessarily be kept confidential, all deliberations on these should be in public, and justification / reasoning be given for any actions against an editor.

Accuracy

See: wp:Accuracy which evolved from here.

WP:Notability

Issues

The complex kludge which wp notability (consisting of guidelines, guidelines on how guidelines interact, incorporated policies, wiki-venues, essays, lists, traditions, influence from history, crowdsourcing and more) works most of the time but is so complex that nobody can describe or improve it. There is not even a definition of this "wp:notability" quality that we screen for. Wikipedia:How Wikipedia notability works explains how it actually works. This page should be prominently linked and evolved.

New constitution for WMF or its replacement

Start with the current one and make following changes:

• Jimbo's seat is embedded, rather than subject to appointment by the board.
• In the event of a tie vote, Jimbo gets a second tie breaker vote
• All other board seats are elected by the community, not 3 self-appointed by the board
• Changing the constitution takers a 2/3 vote by the community, not a majority vote by the board.
• Added a bill of rights
• Added: The technical ability (e.g. via control of the servers) to exert ultimate control does not per se grant any additional authority.

Article XI Bill of rights

1. Community members are entitled to due process, commensurate with the severity of a sanction that is imposed
2. Anyone who is a target of a sanction shall be informed of the reason for the sanction

New Page Patrol capacity and backlog

The backlog of unreviewed pages is an ongoing unresolved problem which is getting worse. As of this writing it is 15,000 and rapidly growing. This is the result of hundreds of factors, some good, some bad. Any useful analysis requires identifying core solvable problems or good changes which will help improve the situation. I'll discount some other areas discussed as too small to help (e.g. autpopatrol and bot related) and focus on the core need and goal which is the amount of throughput of by-human NPP patrol reviews. For shorthand/terminology and ballpark I'm also going to use "30 reviewers" for shorthand because currently this is the number of reviewers averaging 2 review completions per day or more, and under best case scenarios, the number we might have reviewing big enough numbers to solve the issues. And I'll use '1 million" for the number of editors, knowing that it can be considered to be more or less than that depending on the criteria used.

General Analysis

It appears that a throughput of about 800 by-human review completions per day is needed. Since maybe a third are just tagged for later re-review, this means about 2,200 reviews a day including those. Attempted reviews where the NPP'er ended up just leaving it it for an expert would raise that number further.

The gorilla in the math living room that affects everything else is that under the best case scenario we have millions of article creators and 30 people doing NPP on what they create.

Analyzing one level down on the throughput side, throughput is the number of active reviewers and the rate of reviews that they can and do accomplish. And one level below that the factors are influx of reviewers that are really going to review, having them get better at it, having them spend more time reviewing, making it easier to accomplish reviews, and retaining them as active reviewers. And of course, making the process less difficult and more pleasant is key to many of these.

Even to get to moderate proficiency requires a wiki-knowledge level which is higher than that of an average admin. For example, to be able to review even 1/2 of the articles that need review requires extensive knowledge of all of the notability guidelines, wp:not, wp:common outcomes and some knowledge of of the undocumented fuzzy wp:notability ecosystems works. To get that to 95% requires far more than that, with medium knowledge of most of the fields covered by the SNG (e.g. to know the sports terminology and structure to interpret the SNG), fluency with a wide range of tools (source-searching, copyvio and translation tools) 4-5 other major guidelines and policies and fluency in searching sources in non-english countries.

If one listens to everybody who comments or complains on what seems to be implied, the expectations of what could be covered in a NPP review are gigantic. The list of all of the tags to apply implies catching every possible problem or weakness in an article. The NPP flowchart is not really as demanding as it appears at first glance, but does included copyvio searches and analysis of the results.

NPP'er satisfaction with and enjoyment of the work is essential to recruiting reviewers, having them do more, and staying around. Even if they ignore the complainers, the unwritten expectations inflicted by NPP itself are so high that nearly everybody is going to think "I'm doing a bad job". And, to add to that it is a painful job once one sees comments and happenings at articles that they AFD and similar places.

Of all of the zillions of possible tasks for NPP, 95% could be called article development including identifying and fixing problems. Mathematically the 95% needs to be handled by the millions of editors, not the 30 NPP'ers. The other 5% that can really only be done by NPP is acting as the gatekeepers on "is it OK for this article to exist?" questions.

WP:Before regarding sourcing inflicts a mathematical impossibility on NPP. For an extreme example to illustrate,in 10 minutes, each of the million editors could create 10 sourceless articles on non-notable people in non-english-speaking countries. Each will require about a 1/2 hour by an NPP'er to comply with wp:before before sending it to AFD. So those 10 million articles will require 5,000,000 person hours by the 30 NPP'ers to send them to AFD. If they each average 5 wikihours a week to work on NPP, it would take the NPP'ers 641 years to "properly" review what the million editors can create in 10 minutes. Mathematically, the only solution to this doubly lopsided situation is that including needed wp:notability-related sources (thus confirming that they exist) needs to be the job of the 1,000,00 editors, not the 30 reviewers.

Solutions

1. The general expectation within NPP needs to change to the only responsibility of the 30 reviewers being to be gatekeepers on the "Is it OK for this article exist?" questions. Everything else it's cool for NPP'ers to do but will be considered to be "above and beyond" their responsibility. And we need to convey this whenever it comes up. The change is mostly psychological; don't change the flow chart, just interpret it in this context.
2. It's ridiculous that our main tool (curation) here 50% doesn't work and isn't documented. We need to demand that WWF needs to stop their unneeded self-assigned ivory tower tasks and provide us with a working, maintained. documented tool. Until then, see below that work-arounds need to be covered in training.
3. Finding and putting in notability-required sources needs to be considered a main part of building an article, something done by the million editors. Specifically wp:before needs to be modified to reflect this. Supplying wp:nogtability required sources (and thus determining whether or not such exist) is the job of the million editors, not the 30 NPP'ers (or the AFD folks) NPP'ers can get that done. Until then, we need to implement that unilaterally during this emergency condition. When wp:notability requires sourcing, that we go by the sources that are IN the article.
4. Training here needs to get tweaked. The first stage needs to say to learn and get good at wp:notability guidelines, wp:not and the other ~6 guidelines related to the core task and come back after that. It's too much to try to teach that here. The next stage needs to include a lot of mission helpful getting started tips. (like how to use curation, when not to use it, and (simple as it is) how and when to use twinkle. And later, when reviewers are approaching near-rock star fluency, for the rock stars to give them tips on (off wiki) on practical decisions made to operate at rock-star rates.
5. Implement other items within NPP top help in the other areas identified in the analysis.

Numerical data only article

wp:not needs expansion in this area.

Wikipedia is an encyclopedia, not a presentation of numerical data.