[ Main article: Wikipedia:Relevance ] — the last version for the focus of this talk was REL4.2

Guideline definition

A guideline is any page that is: (1) actionable (i.e. it recommends, or recommends against, an action to be taken by editors) and (2) authorized by consensus. Guidelines are not set in stone and should be treated with common sense and the occasional exception. Amendments to a guideline should be discussed on its talk page, not on a new page — although it's generally acceptable to edit a guideline to improve it. Disputes over the wording of a guideline are resolved by considering and discussing objections and counter-proposals and coming to agreement, often using compromise language; such a dispute does not "suspend" the guideline or "turn it into" something other than a guideline. People are sometimes tempted to call a vote on a guideline, but this is a bad idea because it polarizes the issue.

Originally at: User talk:Father Goose/Relevance/Archive 1

A path of infinite regression

Father Goose, I just don't see this working:

Relevance can usually be demonstrated by answering one of the following three questions...

for the goal of assisting editors in dispute over whether of not something is relevant. I predict that when in a dispute, answering these questions will just shift the dispute into other issues of relevancy but will not help solve the dispute. —WikiLen 06:54, 3 July 2007 (UTC)[reply]

For example:

Maria Diamond is one of 24 featured speakers in The Secret (2006 film). She is in the film as a feng shui expert. For this example, assume she is also an expert on vision boards — an expertise of hers not mentioned in the film. However, the film talks about both feng shui and vision boards. Which should the article on the film say:

A: Maria Diamond, an expert on feng shui and vision boards, ...
B: Maria Diamond, an expert on feng shui, ...

The correct answer is (B) and it has a simple reason: "film is not using her vision board expertise."

Unfortunately, it would be all too easy to make complicated and wrong arguments for (A) by using the proposed policy in ways that are not relevant:

False argument an editor might make	Incorrectly uses proposed policy by claiming	Reasoning an experienced editor would need to make
Since vision boards are in the film, it impacts the film to know that Ms. Diamond has expertise in such.	impacts subject	spamming for Ms. Diamond
Ms. Diamond's expertise is what distinguished her for inclusion in the film. As such, all of her expertise relevant to the article's topic should be mentioned.	distinguishing trait	expertise on vision boards distinguishes her from other experts in the world at large, but not from the other experts in the film
Vision boards are fundamental to the topic of the article, so Ms. Diamond's expertise in it belongs in.	fundamental property	Ms. Diamond's expertise in vision boards is not fundamental to the film — not even mentioned

If an editor's fails on correctly determining whether something is relevant, he or she is also likely to fail at making an argument for it that is relevant. In the end, it will still be the voice of reason that wins the day — that pulls an editor out of regression into nonsense. Either the editor hears the reasoning of the more experienced editor or the editor is forced to yield under the weight of many editors reasoning against him/her. And of course, I may be regressing into nonsense, but it is you that will rescue me, not some policy. We have is no external reliable source to tell us what is relevant. —WikiLen 06:54, 3 July 2007 (UTC)[reply]

Sadly, there is an infinite regression problem on the definition side of things too, i.e.:

Declare relevance as: "must be about the topic of the article"
Define what "about the topic..." means
Define the terms in above definition
Define what you mean when you define the terms
And so-on...

I overlooked one level in my hypothetical definition-regression, the top level: "must be about the topic of the article". I have revised the definition list, above, to include this missing piece. I am contemplating adding a Relevance discussion topic on "What is so bad about a simple definition" — seems the simple definition is: "must be about the topic of the article". —WikiLen 18:16, 5 July 2007 (UTC)[reply]

All language can be open to interpretation. That's no reason to give up on using it, though. The current draft contains at least the first two bullet points listed above. Is it hopelessly inadequate or ambiguous?--Father Goose 21:30, 5 July 2007 (UTC)[reply]

In general, a policy to settle relevancy disputes, if not crafted carefully, is going to make for more work in resolving disputes, not less. This not to say that your obsevations about relevance are not correct. They look very good to me and have educational value but relevancy policy should be ignored—except for extreme cases—in favor of the voice of reason when resolving a dispute on relevancy. —WikiLen 06:54, 3 July 2007 (UTC)[reply]

Right. Which is why I tried to avoid definitions in favor of (hopefully) more practical approaches to evaluating relevance.

Wikipedia:Ignore all rules holds true everywhere. I've been aiming the proposal at that voice of reason -- and offering it ammunition, for both deletion and inclusion rationales. This doesn't cancel itself out: the proposal offers ways to demonstrate the relevancy of something that might not appear relevant initially, and if no convincing rationale is presented, "editors are within their right" to delete the material. The onus is on the "includer" to convince other editors, and the specified criteria are a good toolkit with which to demonstrate relevance (or fail, by using them in an unconvincing argument).

The guideline says right up front that relevance is ultimately gauged by editors (I just added "collectively" for more specificity). It doesn't give an individual power to sweep aside others' views. But I see what you're getting at here: a tendentious crank might latch upon the "trio" and use them in a distorted way to argue for inclusion of his pet factoids. I'll change the wording to stress "use the criteria to convince other editors" more.--Father Goose 17:19, 3 July 2007 (UTC)[reply]

Yes, Father Goose, I find you get my point and I expect you disagree with some of it. Isn't "a tendentious crank" in all of us at times? That is, can't one's emotions cloud the reasoning of even the best of editors? I wonder if you way under-estimate how many extended disputes are happening because one or both editor's reasoning ability is clouded by emotion (or do I over-estimate). With "Relevance" there is no external source to clear the cloud and a written policy is too amorphous a thing to be successful at doing that — can't address every possible mis-understanding. It is only other editors that can pull them out, by holding firmly to what is the right edit. —WikiLen 18:16, 5 July 2007 (UTC)[reply]

I've found, even when I was "on a mission", that the more I had to explain my position, the more clarity about it I gained in my own head, to the point where sometimes I changed my mind. There are also times when I've managed to convince other editors of my points, or they've convinced me of theirs.

Well spoken. This is Wikipedia working well and a relevance policy, hopefully, could make these discussion more efficient. My concern is that, if not careful, a relevance policy would make such discussions less efficient. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

There are still times when reasonable, rationally-acting editors fail to agree on some points, and other times when unreasonable editors fail to agree with others. The last group is the "cranks", and it includes people who refuse to discuss things, or who fail to convince others but refuse to yield to a "consensus" position which opposes theirs. Their actions hold little weight on Wikipedia: if they continue to enforce them in the face of opposition, they tend to get banned.

Agreed. This is also Wikipedia working well. No problem to solve here with policy (do cranks even read policy). No policy is going to transform cranks or make cranks go away and anyway, editors already know how to deal with them. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

What I've been attempting to do with the guideline is first to document the "consensus position" on relevance, where it does exist (via WP:NOT, WP:BLP, and other policies, and some elaborations of my own which I hope are found to be agreeable to most); and second, to provide a few "acid tests" which can be used during discussions of relevancy. This is similar to what WP:Notability and related guidelines do, and it at least does something to pare away WP:ILIKEIT vs. WP:IDONTLIKEIT, which does nothing to resolve disputes. I'm not claiming I've found any "perfect" expression of relevance, but I think I've managed to put together a solid and practical middle ground on the issue.--Father Goose 21:30, 5 July 2007 (UTC)[reply]

It is the "acid tests" that needs attention. It is what I have trouble seeing working. What counts is whether the "acid tests" shorten the discussion between editors. I expect not, at least in the three-test form you are proposing. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

By the way, why is the three-test form more efficient — for resolving disputes — than the simple form: "Is it about the topic of the article?" —WikiLen 05:38, 6 July 2007 (UTC)[reply]

The guideline also advises (or will; I need to add the "context" section back in) that explanations on talk pages aren't enough; context for relevancy should be put in the article itself. I've been evaluating the guideline through a number of test cases, and the one Mlewan brought up at WT:REL -- rose -- was a nice one. Putting the guideline into action (theoretically, anyway; I didn't perform the actual edits), I found that providing that context had a transformative effect. This is, again, similar to what Notability advises: some articles can be salvaged simply by more clearly stating the notability of the subject; some "trivia", or lists, or other seeming cruft can be salvaged by better explaining how it relates to the subject.--Father Goose 21:43, 5 July 2007 (UTC)[reply]

This I whole-heartedly agree with. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

I suspect you are inclined to think educational material can also be useful when resolving—adjudicating—a dispute. I differ on that. My inclination is to push for a clear separation, in the policy, between what educates and what adjudicates. Stuff that helps adjudicate a dispute will, I suspect, not teach much, but most importantly, stuff that educates an editor seems likely to complicate the task of adjudicating disputes — so... I advocate keeping the two from being confused with each other and clearly stating in the policy that the educational portion is not expected to help adjudicate disputes. —WikiLen 18:16, 5 July 2007 (UTC)[reply]

I'm afraid I don't understand this statement at all. Can you give me an example?--Father Goose 21:30, 5 July 2007 (UTC)[reply]

I am still of the mind that the policy on relevancy needs to be divided into two separate sections or articles; one focused on improving the editor's skill at selecting relevant material for an article and another that gives aid to editors in disputes over relevancy. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

To teach an editor how to identify what type of relevancy a fact/item has is useful in my estimation — what your proposal does well. It will help the editor to write good Wikipedia articles. However, given two editors who have both been so educated, I don't see the three-test "acid test" helping them when they have non-trivial conflicts. Because of the complexity of relevancy, it will just give editors more things to digress on or get confused about — if they use the policy. (They may find the policy too complicated.) The charm however in you approach — as an educational tool — is that it may reduce the number of time consuming edit conflicts. But if at the same breath, we tell all editors to use this three-test "acid test", when resolving relevancy disputes, I fear the remaining non-trivial relevancy disputes will be much more time consuming. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

As an example, teaching two people that defensive driving is the best policy when driving a car, will make them better drivers. However, should they crash into each other in their cars, their driving skills will give them no wisdom or tools for judging who was at fault. In the case of Wikipedia, it is our passions we are driving. When you apply your "acid test" to what you read, you find it works fine — that's solo work with one's own passions and no need for words. I need to see your relevancy tests applied in an actual conflict by the parties of the conflict. Your tests are very technical and take some effort of focus to use. I think editors will find it hard to put into compelling words how something meets — or doesn't — one of the three tests you propose even though they may be able to see it for themselves. —WikiLen 05:38, 6 July 2007 (UTC)[reply]

Okay, I understand now. Yes, the three-question approach may fail to gain traction. It's not field-tested, except mostly by myself, and even I've found it can take some pretty deep analysis sometimes before I can figure out how to apply it sensibly. But I don't think this is necessarily a failing of the "tests": it may simply be representative of how tricky it can be sometimes to articulate why something is "relevant", even though you're sure it has merit.

I've filled in the "Impact" and "Distinguishing traits" sections, with what I hope are passable generalizations of how to apply those criteria. Even if the proposal doesn't get promoted, I feel I could use its advice right now to make a number of cogent arguments about the relevance of specific bits of information.

One of the big unresolved issues on Wikipedia regarding relevance is what MangoJuice called "connective trivia" at WP:HTRIV. To me, the perfect example of a "relevant connection" is the use of The Blue Danube in 2001: A Space Odyssey. It's a pop-culture connection, but a very significant one; it's concevably the most iconic use of classical music in any film, ever. And yet, when I went to check the Blue Danube article today, I found that mention of 2001 was purged, along with dozens of other "Blue Danube in pop culture" references, months ago.

Now, this is nuts, and it probably wouldn't be hard to convince even the purging editors to retain at least that reference. A ready-made argument is in the "impact" section: Causing the subject to come to public attention (i.e., increasing its notability). Maybe all the proposal needs is more of those "readymades" spelled out. It seems to me that it's mostly complete right now, but with more testing, I'm sure more stuff will come to light.

I'm gonna add one more section -- a revised "Connections to other subjects" section -- then post it back at WP:REL. Thanks for your feedback so far.--Father Goose 08:49, 6 July 2007 (UTC)[reply]

Is it about the topic of the article?

By the way, why is the three-test form more efficient — for resolving disputes — than the simple form: "Is it about the topic of the article?"—WikiLen 05:38, 6 July 2007 (UTC)

Neither of our approaches has undergone any kind of rigorous trial. But it's my feeling your simple form is, well, too simple. Some facts may be "about the topic of the article", but still really unimportant. I'll give you two examples I came across recently:

Unimportant would be a separate criteria. See this below at, Wikipedia talk:Relevance#On policy verses guidance

4. Boundary at micro level (reach into details must be driven by what serves the reader) —WikiLen 06:24, 8 July 2007 (UTC)[reply]

Example 1: "The 1998 film The Siege, starring Denzel Washington, features the WTC in the background."

This is from World Trade Center in popular culture, and there are dozens of equally incidental examples on that page. They were even more irrelevant when they were in the World Trade Center article. The line Incidental connections between subjects (i.e., with no demonstrable impact on either) do not need to be documented anywhere on Wikipedia. is specifically targeted at this kind of irrelevancy.

I like—find it belongs—the list of incidental connections that the Trade Towers have to culture, as listed at World Trade Center in popular culture. It has almanac character that serves the readers interest considering the iconic character the Trade Towers have. It is not often mentioned, but Wikipedia serves as both an encyclopedia and almanac, as mentioned in one of Wikipedia's tutorials: "Wikipedia is an editable encyclopedia (along with some topics that would typically be found in an almanac)". I would argue "The Siege" belongs in the list because the importance of the topic elevates the relevance of what would otherwise be details too trivial to include. Furthermore, I would argue the line in question is more about the World Trade Center in popular culture than it is about the film, The Siege and so belongs in the Trade Towers article if anywhere. —WikiLen 06:24, 8 July 2007 (UTC)[reply]

Example 2: A majority of the permanent flavors offered by the company include chocolate in one form or another, though there are vanilla-based blends as well.

From Häagen-Dazs. Sure, it's "about" Haagen-Dazs, but it's a pretty unremarkable statement. Even if it were actually true (by my count 21/49 include chocolate), it could be shot down with "not a distinguishing trait" because lots of ice cream flavors tend to involve chocolate. It would, however, be relevant to a "Flavors" section -- as part a fundamental description of that subtopic.--Father Goose 05:27, 7 July 2007 (UTC)[reply]

If I were to argue the relevancy point I would say: Is relevant, as it is about the topic of the article, Häagen-Dazs — i.e.: Häagen-Dazs makes lots of ice cream with chocolate in it. I might also argue that, although relevant, it is too trivial a point to include in the article. —WikiLen 06:24, 8 July 2007 (UTC)[reply]

Improve articles verses improve editing efficiency?

[Father Goose, an eureka hit me after working on your last comments above. A focus on efficiency is the heart of what needs to be done and I have realized how to do it. Note: I have also added this section to the talk page for the main article — WikiLen]

I see two agendas in contention for a relevancy policy:

Improve quality of articles.
Improve efficiency — speed up resolution of edit conflicts.

I think the primary agenda should be (2) — improve efficiency and thereby indirectly improve article quality. I propose designing a policy by:

Weighing what impact elements of policy will have on edit-conflict efficiency.
Codifying only what good editors are already doing (empowering their voice-of-reason during edit conflicts).

What good editors do will evolve and the policy can then change to reflect the shift. In short, what good editors do should drive the relevancy policy (resulting policy teaches other editors). A policy that drives all editors is just not going to work. That is, any attempt to legislate from on-high what does & doesn't go into an article is an academic exercise that editors cannot unify around. The relevancy policy-by-example, that good editors have already engendered, approaches the limit of what good editors can unify around — lets codify it. —WikiLen 08:36, 8 July 2007 (UTC)[reply]

I realize this calls for morphing your ideas into being the educational component of the policy. —WikiLen 08:36, 8 July 2007 (UTC)[reply]

I look forward to seeing what you have. I think now is the time to present it. I — and others — should look at it before passing judgement on or proceeding with my "improve efficiency" approach. And I may be pleasantly surprised at what you have done. —WikiLen 08:59, 8 July 2007 (UTC)[reply]

Are you referring to the completed 2.0 draft? I copied it over to WP:REL two days ago. You have looked at it, right? Regardless of whether you still disagree with the "three questions", do you think the other sections are all right?--Father Goose 10:10, 8 July 2007 (UTC)[reply]

OK, I see the new draft. I found the "Relevance in subtopics" section did not seem right — did an edit on it. I find the tone of sentences a bit too passive — a minor point. On the "three questions" isssue, it is not that they don't belong in it, rather in what context they are presented and what recommendations are made about their use. Also, the "relevancy transcends attempts to codify" point seems missing or tucked away as a minor point. I plan to do a major refactoring organizing the policy into two major sections: (1) "On addressing edit-conflicts" and (2) "On improving one's own editing." I expect your contribution to substantially remain present.

Okay, I'll keep my hands off it for a while to see what direction you'd take it in. I'll offer comments as your work proceeds.--Father Goose 03:19, 9 July 2007 (UTC)[reply]

Originally at: User talk:Wiki Len/Relevance

Explanations for my changes

I'll do what I can to explain the changes I made here:

Moved the disambiguation heading to the top; this is standard format (for instance, see WP:DR)
Removed {{policylist}}; this isn't a policy; no other guidelines include it.
Typically... Often... To be relevant... -- these three sentences all say more or less the same thing, so I folded them into one sentence. If something is obvious, there's no need to say it's obvious.
I removed the footnote, which was a mini-editorial. Either say it in the body of the guideline or don't say it at all.
Don't break up your thoughts by using bullets. Bullets should be used for discrete "items", not sentences that could be prose instead.
There's a problem with "content duplicates content"; if the content is duplicated, it's already there, no point to moving it. That error in wording was mine, from REL2; I fixed it, at any rate.
I removed all the intra-page links and shortcuts. Should any part of the guideline be read in isolation? It's not a collection of discrete rules like WP:NOT.
Don't use a numbered list unless you're laying out sequential steps.
emergence of relevant content... Relevant content emerges... -- redundant. I removed the second one.
"not by application of policy or reliable sources," -- how necessary is it to say this?
(serve the readers) -- unless you explain how to serve readers' needs, there's no point to invoking them, especially not as a parenthetical in a section heading.
"balanced against notability" -- What's notability got to do with this?
The three bullet points that follow should be turned into prose.
I changed the two links to your essays from {{main}} into {{see}}; you shouldn't cite essays as the "main articles" for material in a guideline. In fact, I'm not sure you should cite them in the guideline at all, even though they represent your views on the issue.
You link to WP:NOTABILITY twice in proximity; I removed the second link and, separately, sharpened the sentence in which it appeared.
That footnote doesn't explain anything. Even less so as a footnote.
I tried to tighten up the wording of the bullet points... which are not necessary anyway.
I removed the one that outright contradicts WP:TRIVIA. If you're trying to change or revoke that guideline, do it there.

Anyhow, that's my review-cum-edit of REL4. I hope you can make use of some of it.--Father Goose 20:10, 20 July 2007 (UTC)[reply]

Fantastic, a version REL4.1 —WikiLen 08:47, 21 July 2007 (UTC)[reply]