Wikipedia talk:Notability/RFC:compromise/Archive 3

Archive 1

Archive 2

Archive 3

Lists

One of the things that may help might be to define the purpose of lists.

• Lists are useful for navigation (some might claim that dab pages are essentially a type of "lists").

• Lists are useful "breeding grounds" for new articles. (Any time a section of a list becomes developed enough, it can be split into it's own article.)

• Lists are a useful place to merge slow-developing stubs. (It helps provide context to the overall topic that the stub may have been covering, and so making it easier to develop through sources and reciprical information.)

• Lists are useful as a place to indicate a desire for a new article. (Add a redlink and cross your fingers : )

• Lists are useful in helping to provide context to an overall topic or group of articles.

• Lists are a useful organisational tool for displaying/presenting information, and have several ways to do so.

• Lists can "self-organise" through the use of "parent lists" (not unlike "parent cats" as part of a navigational structure. (See the main page for some contents, and indices, for example.)

And these are merely a few of the more obvious uses.

The problem is that most editors don't look for the possible uses of a page, and are merely interested in whether they feel such information should be presented.

Awhile back, Lists of fictional things was deleted simply because several of those commenting in the AfD didn't realise that this was a "parent list", providing navigation to subordinate lists. (I may DRV it in the near future - it's been on my "back burner" for awhile.)

One of the main problems I personally see with lists is a question of whether there are inappropriate intersections in the inclusion criteria of the list.

For example:

"List of characters on a TV series" - too broad, but might be useful as a "parent list" pointing to lists with more specific modifiers. (Another way to look at this is to presume that all lists with those two intersections are a part of this list, but merely are split to separate lists due to WP:Summary style and WP:SIZE.) To "split", select one of the two intersections (either characters, or TV series) and substitute a more specific modifier.

"List of doctors on a TV series" - the original two intersections are still there (fictional characters, and television series), but one of them (Fictional characters) has been modified to become more specific (doctors). The more specific modifier should be an encyclopedic topic and thus would be "notable". ("List of fictional head scratchers on a TV series" comes immediately to mind.) Note that in this case, this list perhaps still too broad, and may be further split.

"List of doctors on ER" "List of doctors on General Hospital" etc. - This time the other intersection criteria (TV series) was modified. And again the new modifier(s) are encyclopedic topics.

"List of secret agents on Heroes with superpowers" - This adds two different modifiers to "Fictional characters": secret agents, and super powers. This is just a case of more specificity.

Though we should beware of splitting lists to become so specific as to only have a couple of entries. (Small, with little hope for expansion: "List of secret agents on Heroes with the power to manipulate fire with their hands".) In which case, an "upmerge" to a less specific list might be appropriate. (No different than when "upmerging" a stub to a list for potential further development.)

And the above should provide a start on helping to establish a "notability guideline" concerning lists.

We simply look for the "core" intersections, and determine of they, or the "further modifiers" each are notable. If not, then the list's inclusion criteria needs to be modified, and the list pruned. If that means that the list becomes empty, or becomes a duplicate of another list, then deletion is appropriate.

Consider "List of blondes on Heroes", for example. We have the intersection of (fictional characters) and (TV series). The TV series may be notable (Heroes), but the modifier to the fictional character (blondes), may not be.

And if consensus determines that "blondes" is inappropriate (trivial) for inclusion criteria, then merely "up"-grade the list. So the list would become "List of fictional characters on Heroes" (since "blondes" was the modifier for "fictional characters" in this case). And since that list (presumably) already exists, then the result would simply be: Merge and Delete.

Simple.

(And now we wait for me to be flamed : ) - jc37 22:39, 8 October 2008 (UTC)

...what? I know you said "simple" but I don't comprehend what it is you're saying. All I see are bunch of completely arbitrary ways of breaking down a list of things that may or may not need to be listed in the first place. Nifboy (talk) 23:55, 8 October 2008 (UTC)

It's only by breaking these down into their core components that we're going to be able to develop any sort of honest consensus on any one specific list. - jc37 21:25, 9 October 2008 (UTC)

Still lost. I like exactly zero of your examples, as they seem more suited for a list on TV Tropes or some such, where you can list anything so long as it meets the criteria. Incidentally, no, I don't like List of fictional doctors (or, for that matter, most of Category:Lists of fictional characters); it's both hopelessly incomplete and so big it's threatening to collapse under its own weight. Some of your proposed "purposes" of lists feel like solutions looking for a problem: we hardly need to encourage new articles, especially in fiction.. Nifboy (talk) 00:18, 10 October 2008 (UTC)

I think we should encourage more articles in every facet of Wikipedia including fiction. Even should we eventually have a billion articles, we should ravenously push for more : ) - jc37 10:22, 10 October 2008 (UTC)

Give me a thousand more well-organized, high-quality articles over a billion articles of indeterminable quality any day. Nifboy (talk) 16:55, 10 October 2008 (UTC)

I think this is a great starting point. I might want to make sure that the criteria of the list itself is not "trivial" (e.g.: not "list of TV characters with blond hair"), and I definitely need a little more clarification about how you might write a list article without reliable third-party sources. But let's see what other people think, particularly those on the "anti-notability" end of the spectrum. Randomran (talk) 18:54, 9 October 2008 (UTC)

So let's assume "List of doctors on ER" and "List of doctors on General Hospital" would be alright - would "List of doctors on Stargate SG-1" also automatically be okay (two major recurring doctors, a handful of minor recurring doctors whose names I don't remember, a bunch of one-off background doctors)? If so, why? If not, why? And we're back to square one. – sgeureka ^t•c 20:08, 9 October 2008 (UTC)

I think it's an issue of carving these lists up in a reasonable way. Maybe we need a better way to phrase is, but the proposal states that a list should not have a scope that is "overly limited". Doctors on ER might be reasonable, but doctors on StarGate would be overly limited, let alone trivial. As an aside, I'd wouldn't want more than one character list per series, unless someone had some sources to show that another list is notable (or even semi-notable). Randomran (talk) 20:15, 9 October 2008 (UTC)

On your last point - By the system above, I would rather state that no character list per series should be created unless justified based on development (which is broader than just stating notability). So for example, the "List of doctors on Stargate SG-1" could be treated as a "split" (even if it was merely a creation), and then can be (re-)merged to "List of characters on Stargate SG-1", for further development. Essentially, any list may be potentially split or merged based upon development. - jc37 21:25, 9 October 2008 (UTC)

I think that's a fair point: if two lists should be merged together, there will be a merge. If we have to remove WP:NOT information in the process, then we will. Still, it would be nice to avoid really dumb disputes by explaining what kinds of lists are inappropriate. (Or, alternatively, by having a reasonably high list-standard, but making a few discrete exceptions for things we know are appropriate.) Randomran (talk) 21:49, 9 October 2008 (UTC)

I think the challenge for lists is to create a set of inclusion criteria that do not lead to them failing WP:NOT and other Wikipedia policies. The example given, "List of doctors on Stargate SG-1" would probably fail WP:NOT#DIR, so I don't think jc37 has thought through the implications of his statement that a list is "justified based on development". In the absence of any alternative inclusion criteria, I beleive that compliance with WP:N is defense against failing WP:NOT, unless someone has developed an workable alternative. The proposal made by Erachima is not much better, as based on the idea that a list must be "Related to a notable topic", but this is a classic example of a truism, and as such is too broad to have any practical application because you can't disprove this statement. --Gavin Collins (talk) 11:28, 10 October 2008 (UTC)

Right now it is impossible to write down a full and exacting list of what are allowable or disallowable lists, because there's so many different types that we cannot categorize them. Pending global reassessment of these to make sure there are no problems, there are some that are accept, including lists of episodes and lists of major and minor/recurring characters. There are also ones that are unacceptable: lists of fictional topics grouped from several different works of fiction (eg, list of fictional doctors; this doesn't prevent the use of categories to do this), lists of one-off/cameo characters in a work, etc. We can define the white and blacklists for these, but there's a huge middle ground that I don't think we can qualify yet. What do we do about that? We do a watch-and-see approach: such lists, when put for deletion, should be tracked to try to qualify what the list is: once we have enough to generalize that such a list is generally not allowed or is always allowed, we can black/whitelist it, respectively. This will allow us to slowly chew away at the vague area, though there will always be one like this. But by defining explicit cases we know are acceptable and those that never are, we can provide the right type of examples of what lists are appropriate, with the flexibility that a guideline should be. --MASEM 17:20, 10 October 2008 (UTC)

What is possible is to categorise lists between those that demonstrate notability, and those that don't. Clearly those that don't are deleted, but a notable list would never be deleted. --Gavin Collins (talk) 17:28, 10 October 2008 (UTC)

Note that the category system is designed as a user interface for navigation purposes, not to replace content. Technically we should be able to remove the category system and still have an encyclopedia. (Granted the user would have to click more links when searching for information : ) - jc37 02:11, 11 October 2008 (UTC)

We're talking about all non-notable lists. Notable lists are immediately allowable (save for any other WP:NOT issues). --MASEM 17:33, 10 October 2008 (UTC)

Obviously, lists that abide by WP:N and WP:NOT are acceptable. But the RFC reveals a consensus that there is such a thing as a "semi-notable" list: a compilation of non-notable articles that creates a semi-notable, encyclopedic list. Our job is to figure out what a semi-notable list is. Which lists are allowable? One approach is to come up with a general rule that can be applied in a common sense way, the same way that people try to apply WP:FUC or WP:NOTDIRECTORY with some reasonable differences in interpretation. We might already have it, in fact: Wikipedia articles are not non-encyclopedic cross-categorizations. That would let us get rid of "list of Xs in Y". Another approach for defining appropriate/inappropriate lists is to let SNGs come up with a "whitelist" of allowable lists. But even that would need qualifiers so that someone can't create a 3 person wikiproject with some ridiculous rule. Randomran (talk) 18:56, 10 October 2008 (UTC)

Alternate proposal concerning lists

FWIW, I particularly like User:Erachima/Inclusion (stand-alone lists) in regards to lists. This would leave the list aspect resolved and the SNG/GNG issues can then be taken on. — sephiroth bcr ^(converse) 18:31, 9 October 2008 (UTC)

Any comment from anyone here (aside from Randomran :p) on whether User:Erachima/Inclusion (stand-alone lists) is a suitable compromise on the issue? — sephiroth bcr ^(converse) 21:29, 10 October 2008 (UTC)

Overall, it looks like a decent start. The only place I am uncomfortable is

"Non-trivial" means that the entries within the list are significant to the topic of the list. The significance of individual entries to the topic should be attested by reference to reliable sources. Depending on the topic, these sources may be secondary or primary, but see below.

. I think that particular aspect needs to be tightened up, although I am not sure how. How does a source attest to the importance of the items? Objectively, how do we distinguish between List of characters in Grand Theft Auto 76 vs. List of weapons in Grand Theft Auto 76 vs. List of Street Gangs in Grand Theft Auto 76 and List of autos in Grand Theft Auto 76? What different kinds of things would these sources have to say about each of these things to make one suitable for inclusion and one not?—Kww(talk) 21:49, 10 October 2008 (UTC)

I've mentioned this on Erachima idea's talk page, but I think to help clear up the vagueness of "related to", a long term solution is to define what topics in any field are considered to be worthy of inclusion in WP (inclusion != having an article; a topic must have notability to be given a full article dedicated to it) (User:Masem/Inclusion Guideline is my draft of what I think we can take away from the RFC). What can be defined as being included per field is what are elements that make appropriate lists. So, for example, in a FICT approach, every episode of a notable prime time series should be included in WP somewhere, simply because it is a logical extension of what WP can be when it comes to talking about television shows. The bulk of these lack notability, so these aren't going to be articles, these will constructed into lists. Now, I'm not saying we have to go there first, but if we take approach of how to construct lists from the standpoint that lists are composed of elements that we want to enumerate because we are a broad work of coverage, then we have a repeatable process we can take to other fields. --MASEM 23:01, 10 October 2008 (UTC)

• The idea that list is good if it is "related to" is a truism; since everything is "related to" everything else, it is too broad a criteria to be applied to list inclusion. I don't think a better set of inclusion criteria will be found other than WP:N, which is aleady the standard for Lists of people. --Gavin Collins (talk) 09:22, 11 October 2008 (UTC)
  • That's a pretty high standard, and that standard misses the problem. Okay, Michael Tyson, Michael Jackson, and Michael Jordan are all notable persons. But most Wikipedians would think that a standalone list of Afro American celebrities named Michael would be unencyclopedic and inappropriate, even if the comparison has been raised by Notorious B.I.G. in Victory. I'm convinced we should be able to deal with these inappropriate lists with some kind of list inclusion standard. In fact, I think WP:NOTDIR gives us a great starting point: no non-encyclopedic cross-categorizations. Combined with User:Erachima/Inclusion (stand-alone lists), we should be able to come up with something that works. Even if it requires a little bit of deference to the community to interpret the gray areas. Randomran (talk) 15:57, 11 October 2008 (UTC)
    • The only grey area is which inclusion criteria to apply. Opponents of WP:N still need to come up with alternative or supplementary inclusion criteria, otherwise WP:N is likely to remain dominant by default. --Gavin Collins (talk) 08:07, 13 October 2008 (UTC)

Threading the needle pragmatically

Given that the underlying discussion here remains primarily about fiction articles, what if we, instead of making another attempt to create a policy-level consensus and hoping it actually gets acted upon by the community, approach it pragmatically? After all, there's something approaching support for the view that SNGs are relevant. Why don't we, based on the actual deletion practices, try to frame a SNG for fiction? Here's two very, very quick stabs at the two big parts of that:

Episodes: In general, Wikipedia permits articles on episodes of television series. Such articles may consist of a short plot summary, a cast and crew list, and available information about production or reception. Articles must conform to WP:WAF, and must not contain speculation or interpretation unless it is sourced. Articles that are overwhelmed with speculation, or that have excessively long plot summaries are often deleted.

If it is possible, it is often preferable to cover multiple episodes via a list article instead of having individual articles for each episodes.

Characters: In general, Wikipedia permits articles on major characters of fictional works. Such articles may consist of a short fictional biography, and available information about the development of that character or reception of the character. Articles must conform to WP:WAF, and must not contain speculation or interpretation unless it is sourced. Articles that are overwhelmed with speculation, or that have excessively long biographies are often deleted. Articles about fictional characters who are not deemed "major characters" are generally deleted or merged.

If it is possible, it is generally preferable to cover characters either in the main article for the work of fiction or in a list of characters articles. Creating individual character articles should only be done when the alternatives are not feasible.

These, I think, capture practice. Neither is what I would prefer - but I think both describe what actually happens on AfD, and as such come closer to a practical definition of the consensus on this matter than any principle-based statement can.

What do people think? Phil Sandifer (talk) 15:00, 8 October 2008 (UTC)

• I don't think you approach to episodes will work, as without reliable secondary sources, the articles and lists you describe fail WP:NOT#PLOT. If we were to allow such articles, what would there be to stop TV producers having all their products listed in this way in order to increase sales? I think this approach would result create episode spam.
  As regards characters, you can't list every character in this fashion as this fails WP:NOT#DIR anmore that you can list non-notable people. I think you are ignoring the fact that reliable secondary sources are the best defense against deletion for failing WP:NOT. --Gavin Collins (talk) 15:15, 8 October 2008 (UTC)

• They may fail WP:NOT#PLOT, WP:N, and any number of other things. But the above still seems to me to describe what we actually keep on AfD. One can stand upon what the rules say indefinitely, but on this point we seem to have open jury nullification. Phil Sandifer (talk) 15:40, 8 October 2008 (UTC)

Another observation that hits me as I look at AfDs - and this is not based on thorough study at all - but it seems to be the case that video game characters and the like are unusually likely to be deleted. Also somewhat more likely to go are characters in cartoons and children's' television.

I'm hesitant to try to craft a policy around such specific subject basis, but my gut says this is non-coincidental. Video games, by their nature, have fairly minimalist narratives. (I'm in fact unconvinced that treating them primarily as fiction is even appropriate) Children's television, similarly, is generally more lacking in narrative complexity than adult television. There seems to me, looking at how AfDs swing, a general correlation between substantialness of narrative and inclusion. I'm not sure how to spin this into a description of standard practice, however. Phil Sandifer (talk) 15:08, 8 October 2008 (UTC)

I think we might be jumping the gun. This is bigger than WP:FICT, and we need to understand all SNGs in general first. But I definitely think the RFC supports the idea that certain classes of articles have a lower standard of inclusion. I can't say I have a lot of experience with episodes. But having some experience with characters, lots of non-notable characters are deleted or merged to a list. I think that's consistent with what this RFC is saying too: a somewhat lower (how much lower?) standard for keeping a list, with subject specific guidelines having a somewhat lower (how much lower?) standard for keeping articles in general. I'd like to think that lower standard is still based on something objective and independent, but maybe those indicators are reliable in a different way than what we'd expect for facts about politics or history. I think that's reflected in the fact that video game characters and cartoon characters are more likely to be deleted, although I'm not sure what the underlying principle is. I'm convinced there *is* an underlying principle, though. Randomran (talk) 15:11, 8 October 2008 (UTC)

I think the principle gets back to a key aspect of notability - though we may have evolved it into a strange term of art, at the end of the day, it still means "important." And I think we still get a lot of AfD votes based around views of importance instead of sourcing. Which is unsurprising - remember that WP:N evolved from the SNGs as a sort of general case. But it was still always an attempt to define importance. And on AfD, at least, you haven't lost sight of the goal there. Phil Sandifer (talk) 15:40, 8 October 2008 (UTC)

(EC)That's what the proposed FICT offered (maybe some slight languages differences), and as I've noted had 50% of the !votes supporting it. I think this RFC helps to point out that those 50% of !votes are in line with consensus, and if we were today to try to pass the same FICT on the foundation of the RFC results, it would likely clearly gain consensus.

What I'd rather do, however, is instead of carving out FICT and then deal with WP:N is to figure out how to make an approach to lists in the general sense, allowing FICT to fall within that scheme. I've pointed out my approach on inclusion which is separate from notability; that we should cover a broad number of topics but recognizing that 1) we still are avoiding indiscriminate inclusion and 2) when a topic is just not notable, it needs to be placed with other non-notable topics of the same ilk (aka list articles). What topic we specifically include and exclude are spelled in SNGs. This applies cleanly not only to episodes and characters, but to sports figures, politicians, geographical locations, etc. etc. This approach generally meets how I see the results of the RFC being reconciled for the best, but its also not the easiest route. The short term is to readdress the SNGs with this approach in mind to make sure that flies ok; once there we can talk the inclusion approach. --MASEM 15:14, 8 October 2008 (UTC)

• Your prosal to exempt list from WP:N might have got support at WP:FICT, but the fact remains it would not work. Lists that fail WP:N would still have to run the gauntlet of WP:NOT at AFD. For instance, lists of non-notable video game characters are particularly at risk as they fail WP:PLOT, WP:NOTDIRECTORY and WP:GAMEGUIDE. --Gavin Collins (talk) 15:25, 8 October 2008 (UTC)
• I just am very unconvinced that the way forward is via the general case. We don't delete the general case. We delete specific cases. And I feel like trying to extrapolate from our tendencies in deletion of specific cases is healthier than trying to legislate a general case. Phil Sandifer (talk) 15:40, 8 October 2008 (UTC)

• I think we can deduce that lead ballons don't fly, both generally and specifically, and so it is with non-notable lists; talking about them in abstract terms does not mean the practical implementation of this proposal will work any better than trying to fly a lead balloon. --Gavin Collins (talk) 15:46, 8 October 2008 (UTC)

• I... don't know what to say to this, given that this proposal is based on looking at practical implementations on AfD. Should I just point out that lead balloons float]? I mean, this view is based on looking at AfD, looking at this RFC, and judging what we actually do. Declaring "It doesn't fit the policy" is pointless. It obviously fits the consensus - or, at least, it fits it better than what you're saying, which is transparently and obviously not what we do. Phil Sandifer (talk) 15:49, 8 October 2008 (UTC)

• I'm not suggesting get rid of the general case; GNG still will exist and still is an excellent measure to determine if a topic is notable - it's just that from AFD and the RFC that it is not the only measure for an article, and those we have to figure out how to define better. And I'm not against thinking about the problem in terms of FICT and fixing that first (as that's the area this all started), but isolating improvements to FICT and not thinking how an equivalent approach impacts other fields is going to end us right back to overall complaints. You're right that I think we can assuredly say that lists of episodes of notable shows and lists of major characters of notable fictional works are appropriate; now, can we devise a generalized approach to lists that that says the same thing for people, places, and other aspects? If we do it this way are we giving fictional elements special allowances that other fields don't have? (Personally, I don't think so if we limit to episodes and major characters) Maybe what I'm suggesting means that we have to normalize all the current SNGs to state similar language up until their various criteria; maybe there's something else. It is important to note that the RFC does have strong indication that our current block of SNGs are a mess, whether there are too many, or huge difference in standards, and thus that's why I think the route to resolving this, and allowing for discriminate lists, is through normalizing and improving the SNGs as a group, not individually. --MASEM 16:08, 8 October 2008 (UTC)

• I just don't know that I see some inconsistency over two and a half million articles as a bad thing. I'd worry more about complete consistency over all of our articles than the opposite. Phil Sandifer (talk) 16:27, 8 October 2008 (UTC)

On the subject of the video games, I think the project has pretty resoundly rejected lists of "characters" that are simply obstacles you shoot in their weak spot until they're dead (ex). Nifboy (talk) 17:01, 8 October 2008 (UTC)

I think we can pull out a general principle from this quick discussion of fiction (itself a very narrow issue in for this broad Wikipedia guideline). That's that there is evidence that there are exceptions to the GNG in some form, for some kinds of content. And that's consistent with this RFC on issue B:

• SNGs are here to stay (if you look at B1 and B4)
• But SNGs cannot completely override or circumvent the GNG (if you look at B5)
• Thus SNGs are able to craft specific exceptions within reasonable limits (if we can try to understand B2 and B6)

If we look at the fact that some characters are preserved in spite of sourcing problems, there's probably some kind of principled exception going on (with a bit of noise and error in AFDs, which is natural and expected). But from this RFC, we know that principled exception can't be wholesale ignorance of the GNG. Rather, it's that we use a different objective indicator to judge the notability of characters, and source them according to a different but similar standard. I think this might help us to understand how SNGs function in general, before we jump to the next step of (re)crafting our SNGs. Randomran (talk) 06:25, 9 October 2008 (UTC)

• I don't follow your reasoning that "SNGs are able to craft specific exceptions within reasonable limits". Surely a more logical conclusion would be that SNG's can't craft specific exceptions from GNG at all? --Gavin Collins (talk) 08:23, 9 October 2008 (UTC)
  • Maybe "exception" is a strong word, but B1 and B4 shows that SNGs are meant to allow some kind of relaxed standard for certain kinds of content. The best evidence for how to relax that standard comes from the (conditional) consensus on B2 and/or B6. The RFC reveals that people don't want to enforce the GNG 100% consistently. Randomran (talk) 17:46, 9 October 2008 (UTC)
    • And AfD backs that. What do you think of the wording for episodes and characters I proposed above? Phil Sandifer (talk) 17:49, 9 October 2008 (UTC)
      • I'm not sure, but I'd lean towards something more specific. I think it would be easier to make the case that lists of characters and lists of episodes are notable. But that's not to say that characters and episodes strictly need significant coverage in reliable third party sources in order to stand on their own, outside a list. I think there *is* some standard that causes us to keep Underworld (Doctor Who), but redirect Truth be Told (Alias episode) to Alias episodes (season 1). And I think it might have something to do with the sources in the Doctor Who episode, which are not up to the standard we might expect for a political or history article, but might be reliable when it comes to television. I'm just brainstorming, hence words like "maybe", "might", etc. Randomran (talk) 18:10, 9 October 2008 (UTC)
        • Though if you look at Truth be Told when the redirect was made, I think it would fail the test I set up - it could, at that size, easily be merged upwards. (Whereas I suspect, in practice, it could support an article, as I'm pretty sure there are some decent Alias guidebooks published.) Phil Sandifer (talk) 18:23, 9 October 2008 (UTC)
          • I think we need another criteria for "keep" other than size, though. Some objective indicator, even if there's some room for interpretation. It sounds to me like you're saying that a guidebook would be sufficient to source an episode, though, and show that the episode is worthy of inclusion. Randomran (talk) 18:58, 9 October 2008 (UTC)
            • Sure, though I'm hesitant of some of the implications of that - mostly because I think it equates notability with the existence of a market of fans sufficiently rabid to buy a guidebook like that. And I'm really, really loathe to equate commerce with notability like that - it's one of the major hesitations I have about our sourcing policies. I mean, no Star Trek series has had the success that Law & Order has had, but the nature of the fandoms means that there are countless Star Trek reference books, and very few Law & Order ones. I'm uncomfortable with equating that fact to the notabilities of the episodes or series. Phil Sandifer (talk)
              • Question: For a series without the rabid Star Trek fanbase that would make a guidebook marketable... how would you source it? And by what measure would you judge it as notable, as opposed to a show that might not deserve stand-alone episode articles? Let's brainstorm. Maybe there's a better indicator to use, or maybe we will want to use multiple indicators. Or maybe we'll find that it's not about indicators at all. Randomran (talk) 19:19, 9 October 2008 (UTC)

(Going left for sanity) Well, and this is part of why I lean towards a size-based approach - because I think this gets into tremendously hairy territory very quickly. So let me break this into two parts. First is what I think a viable sense of the overall principle is, second is what I think provides some indicators of notability.

In terms of the general principle, looking at the RFC, and looking at AfD, I think there is some sense in which we consider sub-topics and main topics as a gestalt in looking at coverage. We do not allow unrestricted growth of subtopics, but on the other hand, we implicitly recognize that there are some kinds of forking that we do (and others that we definitely don't). This is not a well-defined point of policy, and I don't think this is the time to try to define it, but I do think that our coverage of a work of fiction is taken both as a set of individual articles and as a whole. The goal, then, is to have appropriate coverage of plot, characters, and production within that whole. Now, of course, the magic word there is "appropriate," and we have some genuine disputes on what that means, but we also all (even the most inclusionist of us - and I'm not talking about me) have a sense that restraint is key here, and that less is more. WP:WAF is, I think, a good guideline, and I think we all like it and wish it were better followed.

So what necessitates more coverage? Two things, I think. The first is popularity of the show - fundamentally, we are going to cover Law & Order more than we cover Century City, and with good reason. The second is narrative complexity. Fundamentally, The Sopranos and The Wire are going to require more than Barney and Friends. (This second point, incidentally, is why I think children's shows and video games get the short end of the stick - lack of narrative complexity.) So as popularity and narrative complexity increase, so do articles. And I think that this tendency - and note, I'm marking a tendency here, not a rule - is borne out on AfD.

That said, second issue - indicators. The existence of published literature on the series is a big one, but we have to note that notability and fan devotion are not equivalent. So what else? Reviews - both in online sources and newspaper sources - are a good indicator. (I'm surprised that Television Without Pity hasn't come up, actually - their recaps/reviews are wholly valid sources - they're a notable outlet. I'm not sure they're a reliable source for factual claims, being self-published, but their views are worth reporting, and go a long way towards justifying articles on hundreds of episodes.) But the caution here is presentism - we don't want to declare that episode articles are more OK for Everwood than for I Love Lucy just by virtue of the fact that Everwood is more recent and has Internet coverage.

Another indicator should probably be DVD release. Both because it often gives us episode commentaries - which is a major source of production information - and because it's a sign of cultural significance that is less dependent on the obsessiveness of fans. I'm certainly not going to say that any TV series that is out on DVD gets episode articles (nor that a series that is not out on DVD can't have them), but it is an indicator. Length of run is also an indicator - more in the negative sense than the positive sense (a quickly cancelled show is almost always less significant - though a few (Firefly (TV series)) are not).

I also think that a softening of "independent" will help a lot. I think that DVD commentaries, interviews with cast and crew, and comments by cast and crew should probably be treated as independent sources for episode and character articles. Not for parent series articles, but remember - a fair part of the independent rule was always to stop companies from advertising. Once we've established the notability of the series as a whole, the advertising concern dissipates, and we can probably afford to relax there. I wouldn't say that commentary by cast or crew provides notability for a TV series. But for an episode within the series? I'm more OK with that (and I suspect that AfD would be as well. That is, I think that "This is one of the episodes that got commentary on the DVD release, so it's an important episode" would swing votes on AfD).

I'm harder pressed to come up with good indicators for narrative complexity. But I'm also somewhat more skeptical that they're needed - this isn't necessarily something where a bright line distinction is possible (if there were an objective measurement of narrative complexity a lot of the wars in English departments over canonical literature would have ended ages ago), and furthermore, I think it's something where, although individuals will disagree, collective consensus can be generated fairly easily. That is, while you will get occasional outliers who insist that all of The Sopranos can be covered adequately in a paragraph or two, or who insist that actually we need 1000 words per episode of Full House to truly capture the narrative, in practice I think that a discussion among multiple people is going to instinctively hone in on a viable consensus here, and that it's a workable yardstick. So I'd be content, in phrasing guidelines, to make a note that narrative complexity will often require more substantial coverage, and avoid quantification on that point.

I think it is probably possible to, for episodes and characters, create a guideline that triangulates between significance and narrative complexity and would serve as an accurate predictor of whether an article will survive AfD. I am uncertain that such a guideline would meet principle-based consensus - that is, I think that inclusionists and exclusionists would both find the phrasing unsatisfying on an ideological level - but I think that such a guideline would accurately predict AfD results, which, I think, is the goal here - to find a way to extrapolate what we do in practice into a guideline. Phil Sandifer (talk) 00:22, 10 October 2008 (UTC)

• Thanks a lot for that really thorough and thoughtful response. I think this gives us a lot to work with.
  • Popularity and narrative complexity are definitely issues. I think popularity kind of takes care of itself on Wikipedia, and is especially apparent when you compare WP's coverage of popular culture to, say, African history. Narrative complexity is definitely an issue, and we might be able to apply this principle to WP:PLOT. Three paragraphs on Tony Soprano may detail some truly insightful details, but you should see some of the deleted video game character articles: talk about clothes, height, weight, friggin blood type. Even if we can't agree on what kind of plot summary is suitable, we can probably agree about what kind of plot summary is patently unsuitable. But I think we're getting off topic. I think this would be a great issue to discuss at another time and place. (In fact, I'd implore you to start such a discussion about WP:NOT#PLOT.)
  • So we know that popularity and complexity affect WP:SIZE, and size affects whether we merge or keep a stand alone article. But that can't be everything, or else would keep an extended entry on virtually any aspect of any topic with zero sources, and arguably zero importance. So let's talk about indicators of notability/inclusion.
  • It sounds to me like you think it would be fair to use sources that are less than independent (fan guides, episode commentary), or sources that might be accurate without having editorial review (television review sites that are generally trusted). I think this is highly consistent with what the RFC has shown. The RFC suggests that we don't generally think every spinout episode or character or *anything* is notable (A1/A2), although we are unsure about the principled exceptions suggested thus far (A3/A4). The RFC also suggests that there *is* an objective way to assess notability that comes from relaxing our sourcing standards (B2/B5). Kind of what you were saying, perhaps SNGs can articulate sources that assert notability that are still widely considered accurate ways to verify information, but might have lower standards of editorial review or be less than independent -- without being the primary source itself.
• I'm thinking out loud. But I think this general principle would explain WP:N in general. Once we do that, we could actually write the SNGs that are slightly friendlier to certain types of content. Randomran (talk) 01:08, 10 October 2008 (UTC)

I like basically all of the above, except (ideologically) the independence issue. Phil, generally speaking, pins down how AfD operates on a case-by-case basis, and yes, I really would love to see WP:WAF get more usage than it currently does. Nifboy (talk) 01:20, 10 October 2008 (UTC)

Just a note that my ignoring this thread for the next few days is because I'm on the road with limited Internet access, not because I'm not interested in it. Phil Sandifer (talk) 11:28, 11 October 2008 (UTC)

All right. I'm loathe to separate discussion of narrative complexity from notability at this point, because I think it's a genuine factor in inclusion/exclusion, and we have to be realistic about the fact that notability guidelines are, in the end, most primarily inclusion/exclusion guidelines.

Regarding your point about how popularity and complexity are not sufficient, you are, of course, correct - a third (and obvious) issue arises. Importance. Which seems to me to be demonstrated in multiple ways - some resemble traditional notability judgments. For instance, I think that part of why a DVD commentary would fly as a grounds for inclusion on AfD is that, despite its lack of independence, it's a clear flag of importance for the episode. But there are other ways that go through more pedestrian avenues - there's a general trend to keep "main characters," for instance, without real regard to sourcing, and I suspect similar grounds could be created for series finales and premieres, and other clear landmark episodes.

I think you're very much right here - there's a wide variety of acceptable sources that should be encouraged to establish notability. I'm thinking the way forward might be a two-tiered SNG - one establishes the general principles of popularity, complexity, and importance. The other establishes that the best way to establish these is via sources, and lists possible sources in a manner that is wider than what WP:N, strictly construed, would advocate. That allows for a strong focus on sourcing while still also remaining honest about the fact that certain things are likely to survive AfD without sources based on a presumption of notability. Phil Sandifer (talk) 21:30, 17 October 2008 (UTC)

I'd be careful with the DVD commentary example, at least to some extent. I've seen unimportant episodes get commentary because every episode on a DVD got commentary, but you still can't show any other info exists for the episode. That said, I'm fair game the the use of DVD commentaries as sufficient for notability - they fall in what I call "1.5 sources" that they are technically primary, but (99% of the time) provide analysis and evaluation from the creates from a point in time separated from the work itself, thus almost acting as secondary sources. The fact that a DVD publishing studio paid to have the DVDs made and incorporate commentary for that episode is a indirect indication that the studio thinks it to be notable as well. Mind you, this is to be taken with a grain of salt - such commentary should include post-showing reaction (I know a handful of simpsons episodes have commentary that say "Episode X is celeb's Y favorite" or somesuch, but even a "on retrospective, we were really drunk and this episode sucks" qualifies too), if there is no other reception section, otherwise, you've got just a plot and a primary source development section and it would be lacking. The RFC does suggest that SNGs can interprent what sources are appropriate for notability demonstration as long as they don't go off the beaten path (eg, forum posts are never going to be RS by an SNG), so I think the same end result can be obtained but without trying to bring in the tricky words of "importance" or the like to it. --MASEM 21:54, 17 October 2008 (UTC)

I was trying to avoid focusing exclusively on fiction... but I think you might be onto something, Masem. One important thing that WP:N does is ensure that an article isn't written almost entirely from primary sources. Primary sources don't tell us what is or isn't important on that subject, let alone if the subject itself is important. You can't write a reception section from primary sources. You can't write a development section from primary sources. You only end up posting a summary, with no reliable independent voice to put it in context. But if we look at the consensus around B2 and B6, it's that SNGs can play with sources. I think this might provide our solution on fiction.
I might venture to say that a fictional topic can be considered notable if it has a reliable information about development and reception. Those normally come from reliable third-party sources. But DVD commentary offers some reliable information about development that isn't necessarily from a third-party, which can highlight the significant parts of the episode or what not. And an award from a highly popular website might let you build up a reception section. The website itself might not be a reliable source,but the award itself might be verifiable, and hopefully the SNGs would have a sense of which awards are popular/prestigious enough to be significant in reception. I'm thinking outloud.
Maybe I used bad examples. But I think the issue really is development/reception that is verifiable and objective, without necessarily being from a strictly reliable or strictly independent source. Randomran (talk) 22:04, 17 October 2008 (UTC)

Reasonable Presumption: Grounds for Compromise?

As much as I hate it, I would say the only way to obtain compromise between inclusionists and deletionist is to expand the inlcusion criteria such that it is permissible to create an seperate page in Wikipedia mainspace (i.e. both articles and lists) in the absence of reliable secondary sources if there is a reasonable presumption that the topic or group of topics on that page is notable. It seems sensible, as this proposal more or less reflects current editorial practise.
The main benefit of this proposal is that no page on Wikipedia mainspace would be exempt from WP:N, but additional inclusion criteria would provide guidance as to what is a reasonable presumption, and how long that presumption could be expected to sustained. I oppose this proposal myself, since I believe that the burden of evidence rests with the editor who creates or restores an article, which must cite reliable secondary sources to demonstrate that its subject matter is notable. The presumption that evidence will be found in a future period effectively shifts this burden from the contributing editor to everyone else. --Gavin Collins (talk) 16:23, 8 October 2008 (UTC)

The reasonable presumption crtierion is already in WP:N: "Although articles should demonstrate the notability of their topics, and articles on topics that do not meet this criteria are generally deleted, it is important to not just consider whether notability is established by the article, but whether it readily could be. When discussing whether to delete or merge an article due to non-notability, the discussion should focus not only on whether notability is established in the article, but on what the probability is that notability could be established. If it is likely that independent sources could be found for a topic, deletion due to lack of notability is inappropriate unless active effort has been made to find these sources. For articles of unclear notability, deletion should be a last resort."

This hardly seems a compromise position. Phil Sandifer (talk) 16:26, 8 October 2008 (UTC)

It think it is perfect if we describe how one would arrive at the conclusion that something is likely to have good sources. Wrad (talk) 21:39, 17 October 2008 (UTC)

time to close?

The RFC has been open for a month and a half, and the watchlist notice has been on for a solid three weeks. It appears as though opinion has crystallized. (Although I admit, consensus can change over time.) I think it might be a good idea to close up this RFC. Is there anyone who objects? Moreover, is there anyone who understands the complex subpages that this RFC has become, and is willing to close it up?

(The next step will be interpreting and parsing the RFC. For now, let's leave that as a separate issue. But I think we should try to get a couple of neutral, non-invested parties to read through the comments and truly understand where the common ground is. Let's worry about that after we've closed.) Randomran (talk) 21:18, 17 October 2008 (UTC)

High-schools

Despite the consensus here that SNGs cannot override GNG, in practice, at AfD there's a sufficiently active group of editors that declare all high-schools notable, regardless of references. Any thoughts on reconciling these two issues? VG ☎ 21:45, 17 October 2008 (UTC)

The consensus is that SNGs don't outright override the GNG. But there's also a consensus that SNGs are needed, and allow the GNG to be sculpted around specific content. Two of the ways we might sculpt the GNG are in B2 and B6, which both deal with sourcing. It's possible that many high schools are notable without requiring reliable third-party sources in the exact same way we would for an article on a historic or political figure. Either way, once we have a firm idea of what the relationship is, we're gonna have to look at the SNGs to see if any of them have overstepped. Randomran (talk) 21:54, 17 October 2008 (UTC)