Wikipedia talk:Featured article criteria/Archive 2

From Wikipedia, the free encyclopedia

Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Size

I've noticed that objections of the general form, "This is too long," have begun to be commonplace on the Featured article candidates page. I would favor adding this to "What is a featured article": Size is not a valid objection standing alone. Wikipedia is not paper; if the article's topic is important, the article's size may be commensurately large." What do folks think of this? Hydriotaphia 01:35, Mar 29, 2005 (UTC)

Readability is a very important criteria. The size of an article affects readability due to the fact that people absorb information based on their attention span. When an article exceeds that point it becomes less useful to the reader. Thus people who need a primer on a topic will not be served by an article that takes 2 or 3 times the amount of time as their attention span (the average upper limit for adults is 20 minutes). A summary with a maximum size not significantly exceeding that limit is needed. More detailed treatment of sub-topics can and should be covered in daughter articles (the main article for that sub-topic). This allows readers to zoom into the level of detail they need. Having everything on one page does not allow for this and is thus not nearly as useful to the great majority of readers (who will either want the concise version - lead section - or the version whose size allows them to read it within the time of their attention span). Everybody wins by Summary style since users who need differing levels of detail get what they want. Oh and citing Wikipedia is not paper shows a fundamental lack of understanding about what summary style does - it in fact makes Wikipedia have more, not less, text on topics (lead sections, survey articles and daughter articles on sub-topics) --mav 17:50, 29 Mar 2005 (UTC)

My own view is that we need a minimum (asy 1500 words) and no maximum. Filiocht | Talk 07:45, Mar 29, 2005 (UTC)
I disagree. "This is too long" can be a perfectly valid objection. Regardless of whether the topic is important or not, having a monster article diminshes its usefulness. — Matt Crypto 10:41, 29 Mar 2005 (UTC)
Yes, that is your opinion, but it is certainly not held by everyone and is not currently in the criteria. Many articles have been promoted in the last months that are 60kb or well over. I personally prefer shorter articles in summary style unless there is a compelling reason not to, but again the criteria don't require that. If you want that to be standard we'd need consensus here to add it as a criteria. - Taxman 14:13, Mar 29, 2005 (UTC)
Either way, I don't think we need to include this as a criterion. Can't this just be a judgement call by the reviewers on FAC? — Matt Crypto 14:48, 29 Mar 2005 (UTC)
No, the criteria are all an editor has to go on in writing a featured quality article. They are there to show the agreed upon standards. People can have lots of opinions about what they would want FA's to be, but an editor cannot know what they are or meet them if they are not in the criteria. - Taxman 14:54, Mar 29, 2005 (UTC)
While we have some guidelines here (Article size), I don't think we can distill what we mean by "Featured quality" into a set of explicit rules. Ultimately, there's going to have to be large elements of human judgement going into reviewing articles. Size is a very flexible parameter. Some articles seem too long at 32k; others seem OK even at 50k — it's a judgement call. We need the fuzziness of human peer review. — Matt Crypto 15:12, 29 Mar 2005 (UTC)

I think Matt Crypto has a valid point here - there are some articles that will inherently tend to be longer, just because of the material (History of Russia being the canonical example) →Raul654 17:16, Mar 29, 2005 (UTC)

I'd still like to see an upper limit set as a guideline going forward. A logical upper limit is the amount of time it takes a person to read before they start to loose focus. That is their attention span. Going past that point is taxing on the reader - who will likely stop reading the article and thus miss how the description of the topic ends. Condensing all the most important detail and presenting that in a form that is highly readable and can be easily read with retention, is doing the reader a great service. Having daughter articles that go into more detail on the sub topic covered in a section, is also very useful. This is not an either or situation - readers who want the detail still get their detail (a series of sub-topic articles), readers who want a quick survey (~32KB or less sized article) get what they need and readers who what a concise version (the lead section) also get what they need. --mav 17:57, 29 Mar 2005 (UTC)
I'm sorry, but how do we know what a person's attention span is? And assuming, and this is a big assumption, that an average attention span can be agreed upon, why is it necessarily "logical" to set that as the upper limit? How do we know that this person won't come back to the article and read the rest at a later time? How do we know that having to click and read a separate article won't deter people from reading the daughter articles? I'm not saying that your points don't have anything going for them. I'm simply saying that (1) what you say is not a viewpoint that has attracted anything close to a consensus, and that therefore (2) it shouldn't be adopted as the official policy of the Featured Article candidates page. Hydriotaphia 18:15, Mar 29, 2005 (UTC)
See Wikipedia:Summary style. It has references for the 20 minute upper limit to adult attention span. I'm not making that up. Readers need the option to able to comfortably read a whole article in one sitting if they so choose. --mav 20:00, 29 Mar 2005 (UTC)
None of those websites referenced any peer-reviewed studies, Mav, so I remain a bit skeptical. (One of them also said that some studies had shown that the average adult attention span was 30 seconds. If that's so, what should we do—make each article a paragraph long? That would seem to be the logical consequence of your position.) Also, please, please, please, make an argument for why "[r]eaders need the option to [be] able to comfortably read a whole article in one sitting if they so choose." Right now it's just an assertion. I remain more-or-less open to being convinced; but I need reasons. Respectfully, Hydriotaphia 06:31, Mar 30, 2005 (UTC)
20 minutes is an upper limit to average adult attention span (I’ll look for a peer reviewed cite on that - but my point was that I was not just making that up). Those with attention deficit disorders and those with below average intelligence of course will have a shorter attention span. Those with very small attention spans or who have very limited time are served by lead sections - which should be concise encyclopedia articles in their own right. There is no need to then have all the coverage on a topic in one article. Dividing it between different articles is highly useful as is having a mid-sized treatment on the topic that can be comfortably read by the average adult in a single sitting and point to the more detailed treatments on sub-topics. All user types win this way. --mav 21:07, 30 Mar 2005 (UTC)

Maybe a compromise can be arrived at. Mav's rigid rule of an article that does not exceed 20 minutes in reading time strikes me as supremely misguided. I think Matt Crypto is right—that some articles seem too long at 32K and others at 50K seem entirely appropriate. It ought to be a topic-specific inquiry. Perhaps the best rule should be: The fact that an article exceeds 32K is not inherently a reason to object to it. We would have all lost out on a marvelous featured article if Mav's rule had been a legitimate reason to reject, for example, the History of Russia. Wikipedia:Summary style presents one viewpoint, but that viewpoint should not become the official viewpoint of Wikipedia, which should be a latitudinarian organization with latitudinarian criteria. Hence I think my first position, that size should never be a valid objection, is likewise inappropriate. Hydriotaphia 18:02, Mar 29, 2005 (UTC)

Summary style is not that rigid. It clearly says that hastily dividing an article once it triggers a page size warning is not a good idea. It also says that an article should be allowed to grow into the 35 to 45 KB range so that a more natural way to divide the article can be explored. I have not objected to any article in the last few months that is less than 45KB in size. Further, the ones I am objecting to are well above 50KB in size. And history of Russia is uncomfortably long as a survey article. Those people who need a compressive yet more compact treatment are not served by a one-size-fits-all approach. We can have the same level of detail as there is now (and a lot more too) by doing more summarizing and having the detail in the series. But since this would be a new rule, it would only work going forward. --mav 19:57, 29 Mar 2005 (UTC)
Ok, then, Mav, I've got two questions. Don't you think splitting up all articles that go above 50KB in size is itself a one-size-fits-all approach? Would you be comfortable with a standard that said, "The fact that an article exceeds 50 KB is not inherently a reason to object to it. Size is a case-by-case inquiry, and the appropriate size for an article is dependent on its subject"? That avoids all one-size-fits-all problems, I'd think, and it also does not exclude from consideration factors like readability. Best, Hydriotaphia 21:43, Mar 29, 2005 (UTC)
No it is not. What is important here are topics - articles just cover at least part of a topic. Having survey/overview articles serve one type of user. Having a set of articles in a series that go into more detail on certain aspects serve another. Both types of articles are needed to very thoroughly cover a topic without deluging the reader with more detail than he or she wants or needs. That is not at all a one-size-fits all approach. What is a one-size-fits all approach is just having one huge article. --mav 20:56, 30 Mar 2005 (UTC)
Dude, that just doesn't make any sense. Of course it's a one-size-fits-all approach—the one size being relatively shorter articles on relatively more topics. It won't fit the hypothetical reader who doesn't wants detail but doesn't want to have to look at a bunch of different articles—who is interested, as it were, in detailed articles about genera, not shorter articles about species. I think you need to admit that your approach is not somehow "neutral" with respect to size preferences. Best, Hydriotaphia 02:08, Mar 31, 2005 (UTC)
There is no real difference between a person who needs a great deal of detail and who is reading a series vs a user who is reading one huge article that has all the same content - but with large sections vs distributing the content on different pages. This user will just be turning the page as it were in the case of a series. At the same time other users may only be interested in one of those pages while yet others will need a higher-level summary of all the subtopics of the topic. Again, just having the one monumental page only serves one type of user. To me that is a one size fits all approach. --mav 17:46, 31 Mar 2005 (UTC)
My attention span runs to about 500 print pages, or; this is the kind of silly criterion we should not go with. I still see no reason for setting an upper limit, provided the information contained in the article is relevant. I am willing to lose my sugestion of a lower limit as a counterbalance to not imposing an upper one. Declaring an interest, I worked on and nominated an FA that is even longer than History of Russia. Filiocht | Talk 07:40, Mar 30, 2005 (UTC)
If your attention span runs that long then you'd still be able to read a series of articles that go into depth on subtopics covered in the survey article. But having just one 500 page long article and no condensed treatment would only serve you and the few others with the time and energy to read so much. Summary style provides for the needs of a much more diverse set of readers and reader needs. --mav 20:56, 30 Mar 2005 (UTC)
I agree with Filiocht's position. No upper limit should be set. At any rate, that seems to be the consensus. Indeed, the only person I've encountered who wants an upper limit is Mav. I hope you'll reconsider your position, Mav. Hydriotaphia 17:52, Mar 30, 2005 (UTC)
Several others have mentioned page size as a reason to object. Just look at the current FACs. --mav 20:56, 30 Mar 2005 (UTC)
I think you're missing the point. Yes, several others have pointed to page size as an objection. What they have not advocated, however, and what you do advocate, is putting a mandatory cap on featured article size. Hydriotaphia 21:01, Mar 30, 2005 (UTC)
Eloquence did exactly that. The others were not involved in a policy discussion but were instead (as I said) objecting based on size. --mav 21:10, 30 Mar 2005 (UTC)
You're right; I was mistaken. OK, then, so it's you and Eloquence. Otherwise, it seems, no one else has subscribed to your recommendation. Hydriotaphia 21:13, Mar 30, 2005 (UTC)
This policy discussion is still young and very few people know about it. :) --mav 21:18, 30 Mar 2005 (UTC)
I agree with Mav that there should be a mandatory cap on page size for featured articles. Splitting off things makes sense when articles get over a certain point, and while that border may be fuzzy, it's not unreasonable to set a hard limit decently above what we should expect any article to be. --Improv 23:30, 30 Mar 2005 (UTC)

What should that limit be, Improv? Hydriotaphia 23:39, Mar 30, 2005 (UTC)

My view of this is that there should be no hard upper limit, but that there should be a strong expectation that long (50k+) articles are the rare exception, rather than the rule. If we were all maths geeks, we might consider a heuristic like the following: we might expect that a 30k article is too long 50% of the time. At 35k, the odds increase to 75%; similarly 40k -> 87.5%, 45k -> 93.8%, 50k -> 96.9% and so on (halving the distance to 100% each 5k). The details don't matter, this is just an example to give the gist of the idea. I'm not saying we should make calculations like this! Rather, I'm suggesting that there should be increasingly strong arguments put forward for promoting abnormally long articles in proportion to the article's length. However, the decision about whether a specific article warrants an excessively long treatment is a judgment that needs to be arrived at by human reviewers, not by metrics. — Matt Crypto 00:16, 31 Mar 2005 (UTC)

That seems a sensible solution to me, Matt. I agree that we should resist the urge to quantify anything. Perhaps the best thing would be a broad standard—as opposed to a hard-and-fast rule—that could be phrased in the following way:

There is a presumption that a featured article candidate's length is appropriate. That presumption weakens in strength as (1) the length of the article increases and (2) the intricacy, complexity and/or importance of the subject decreases. There is, however, nothing inherently wrong with an abnormally long article. Whether an article's length is appropriate must be decided on a case-by-case, fact-sensitive, qualitative basis.

This seems to me to sum up a reasonable position with respect to this issue. What do folks say to the blockquoted standard above? Hydriotaphia 02:01, Mar 31, 2005 (UTC)
I think it is too early to be talking about policy wording. More people need to add their input IMO. I do like the direction Matt is going - especially the particular numbers he is using as examples. :) The example he gave would suggest that articles above 50KB are almost always too long and that even articles below that size could be too long as well (the longer the article the higher the chance of that - deciding what is and what is not too long for the topic would be done on a case-by-case basis).
I can agree with that even though I'd like to have a harder ceiling set: no new FA articles above 45KB of prose while preferring the max size of articles being below 30KB for most topics and understanding that even that may be too long in some cases (prose is just that - tables, markup, references, external links, further reading not counted toward the size).
Encouraging the use of Summary style and Series to cover a topic in great detail should also be encouraged. No reason why we can’t have a great deal of detail on topics - so long as we don’t deluge the reader with more detail than they want or need at any step of the way. --mav 18:08, 31 Mar 2005 (UTC)
Wow! I think if we must have a strictly defined quantitative upper limit, it really ought to be a good deal higher than 45K. Remember also that one of the existing requirements is that the article be comprehensive. While I would certainly concede that many under-45K articles are comprehensive, I very much doubt that every topic can be covered comprehensively by an article below 45K in size. Hydriotaphia 21:40, Mar 31, 2005 (UTC)
I'm talking about mid-level summaries (survey articles) that introduce the reader to a more detailed set of articles (each of which could also be a survey article). Lead sections are concise articles in their own right and a series of articles that go into detail on sub-topics together form a great deal of coverage on a topic (this is the equivalent of one monumental article). Survey articles fill the gap between those two extremes. And just as a good lead section can properly give a quick summary of an article, a proper survey article can properly introduce and touch on each of the major subtopics in a series.
True, some things are so expansive that you can't expect a top level survey article to introduce all the subtopics of a topic; in those cases a second level series is needed (larger issues would be dealt with first and then we hone in on more and more specific items). So a first level survey article on the history of a very long-lived nation would need to divide the history into fairly major periods - each of which would have its own article. Those articles in turn would divide that part of the history into parts - each of which could have articles of their own. And so on until a topic is very throughly covered. That way a reader never gets more detail than he or she needs and those who want the great detail can still get to it easily. --mav 07:29, 2 Apr 2005 (UTC)

The fact is that certain topics are just too big to summarize under a certain size. History of Russia is of course one excellent example: meaty articles can be, and have been, written on many of the individual periods summarized in that article. However, an article with that title cannot simply say "after the Bolshevik revolution, the Soviet Union was formed (see Soviet Union)". That is an injustice to the topic and shortchanges anyone that went to the article "History of Russia" expecting a comprehensive yet concise treatment of that topic. When an article becomes large, the question should not be "is X kb too many?" but "does this article treat its topic in an appropriate amount of detail as concisely as possible?". I would challenge any of you to rewrite History of Russia into 40 kb, and maintain its value as a summary of Russian history. There are simply too many topics that must be touched on; to shorten the article by eliminating some of those topics is to weaken the article itself. I also wanted to remind everyone that there is a handy little feature called the Table of Contents, which provides an excellent tool for parsing large articles (and even reading them in several sessions). I guess what I'm saying is, the length of long articles should be judged against an estimate of how long an article on a given topic needs to be in order to present that topic comprehensively; not against an absolute standard. - Bantman 22:12, Mar 31, 2005 (UTC)

  • I strongly feel that certain articles dealing with complex events and individuals will sometimes and appropriately be relatively large. This is OK, PROVIDED that there is an appropriate lead section. Long articles require lead sections that let the non-motivated reader get what they want and get out. So the focus should be on demanding really well-done lead sections in articles that are longer. If there is to be some arbitrary limit in size, it should definitely be AT LEAST 64K - but I think that it is a bad idea in general. The suggestion that sub-articles can always do the job is, I think, mistaken. Fawcett5 22:41, 31 Mar 2005 (UTC)
  • I agree regarding the importance of a strong lead section. Indeed, the common three-paragraph limitation on lead sections is a problematic restriction on long articles (not to pick another fight...). In theory, the lead section should be brief (something like one paragraph per 10kb, max) but comprehensive enough to A) properly introduce the article; B) provide a quick summary of the article to those with really short attention spans; and C) allow a reader with a narrow topic of intersest to get a good overview, and then use the TOC to find the section he wants to read. - Bantman 00:13, Apr 1, 2005 (UTC)
Lead sections are akin to concise encyclopedia articles. They are called leads in allusion to leading a person into the article. Many users need something more meaty than a lead section but don't have the time or energy to read a booklet-sized article on the subject. The needs of those readers are important and should be addressed. There is no real difference in the reading experience of a person reading a series of related articles (each full sized and on a sub-topic of a larger topic) and a person reading one monumental article. However, the many readers who need a higher-level summary of the topic will not be well-served by the single monumental article. Summary style provides for the needs of all three reader types by providing an article that is not too long (in terms of average attention, and reader time and energy) and not too short. This is a win-win situation. --mav 23:39, 1 Apr 2005 (UTC)

Why should I read an entire article front-to-back at one sitting? If it's long and well-structured, I can read the relevant sections, easily browsing backward if I'm missing references and forward for more information. If it's split into multiple pages, usually in an ad-hoc way, I may have to think harder to find the information I need.

I think people are neglecting the importance of structure within an article. A well-written article allows me to read only the small part containing the information I need rather than the whole article.

There are certainly sprawling articles that would benefit from being split, but some articles are just better in one piece. There's no reason, if the article is well-organized and sectioned, that I should need to read the whole thing.

Splitting up an article has costs: duplication of information, increased effort required for maintenance of linking and factual agreement, increased effort for readers. It has benefits too, but a hard limit would not permit the balance between these to be considered. --Andrew 07:37, Apr 1, 2005 (UTC)

Many people will need to have a high level overview of the whole topic and will not have the time to come back for another sitting. If you have the time, then great! Read a series of articles on subtopics that are introduced in the survey-level article on the topic. Duplication of information across different articles is fine when part of that is in the higher-level summary (which should only have the more important facts) and the other part goes into detail and provides more information on that sub-topic. This reinforces the main points. Duplication within an article should really only happen between the lead and the body and where appropriate to establish context. But the larger the article the harder this is to do. --mav 23:39, 1 Apr 2005 (UTC)
In theory that is more or less acceptable. The problem is when you start to define what "survey-level article" means, and how large such an article may be. Because—as this discussion itself shows—there is huge disagreement on what that means, a hard upper limit on size is, I think, a really bad idea. Anyway, as Fawcett5 and Bantman cogently argue above, a complicated web of articles on subtopics is rendered duplicative by a well-structured introduction. What we really ought to be talking about is getting rid of the 3-paragraph limit on the size of introductions, and more importantly, simply demanding more out of introductions. Hydriotaphia 01:49, Apr 2, 2005 (UTC)
A mid-level of detail is needed between lead sections (which ideally are concise encyclopedia articles in their own right) and a series of closely related articles that go into detail on the sub-topics of something. See below for more. --mav 02:28, 2 Apr 2005 (UTC)
I strongly disagree with the notion that certain articles "have" to be huge. That History of Russia should be allowed to be bigger than Infinite monkey theorem is fine with me, but anything over 40k is in my opinion a weakness. A really good article is one that can cover summaries of very general topics and leave the more specific details to seperate sub-articles. There's nothing stopping us from summarizing facts except ambition. Tolerating 40k+ Features Articles is in my opinion setting the standards for our quality of writing far too low. Peter Isotalo 07:35, Apr 2, 2005 (UTC)
Exactly. One of the most prized skills an author can have, and sadly one of the rarest authors possess, is the ability to correctly summarize so as not to tax the reader's time or energy. This is difficult due to the fact that the author must determine what is most important to his/her audience, how much detail is appropriate for the reader, and to prioritize where that information should go. Doing that correctly is a great service to readers and it is a skill we should be encouraging by promoting the use of summary style. At the same time those who need more detail can and should be provided a series of articles on subtopics. --mav 07:52, 2 Apr 2005 (UTC)

I support a flexible policy on article size. Flexible means that the limit should be a range, not a fixed number, e.g. 32K-55K, where common sense should be applied to determine the appropriate length. As I've already stated on Wikipedia:Featured article candidates/History of South Africa, I believe that we should strive for a reader-friendly encyclopdia which lets you zoom into as much detail as you need. Some limit is obviously needed, as otherwise articles will grow into book length soon enough. We might as well make that limit one which acknowledges the heterogenous nature of our audience.

A 14-year-old who has heard about a particular aspect of a subject in school and wants to get a more general introduction will be well-served with a 35K-40K overview in most cases. If we swamp them with information, we will lose that reader, and he or she will look that information up in a concise encyclopedia instead.

A college student who wants to write a paper about a particular aspect of the subject while providing only circumstantial information about the general topic will want an overview with the ability to zoom into the aspects of the subject that interest them. A researcher or writer will want to have full-length information on the topic.

Perhaps we could make smart use of {{transclusion}} to offer readers a full-length article/book on one page if they desire it -- essentially a skeleton page that merely transcludes the relevant sub-articles. As an online encyclopedia, we have the ability to customize our content in greater ways than a dead tree product; we should fully use the possibilities of wiki-based hypertext to our advantage.

My suggestion is to first strengthen this policy on Wikipedia:Page size, and to then make it a requirement for an article becoming featured. To support the argument, I think it would help to have a "reference implementation", so to speak, an article which makes excellent use of summary style, and perhaps also offers the above mentioned full-length version. Do we have such an article already?--Eloquence* 10:40, Apr 2, 2005 (UTC)

So long as each of those transcluded parts can also act as articles in their own right, I don't see any harm (other than it might be a bit ugly; but we won't know until we try). Their series nav box could have a 'View all' option and direct links to component articles (series nav boxes need to be right of the TOC and/or at the bottom of the page - otherwise they push real content like images out of view). I'd like to call these container articles and and say up front that they should not be considered fully real articles in their own right and they only exist for reader convenience (for those who want a great deal of detail and/or who have high speed Internet and/or who want to print out the whole huge treatment or otherwise have the whole thing in front of them at one time). So these container articles should not be eligible for FA but each component article could.
I can also agree to a max range of 35 to 55KB for individual non-container articles (I'd like it to be 35 to 45KB, but oh well), but a range of 30 to 50 KB for prose might be better (round numbers and page size message starts to be displayed at 30KB, so that a lower limit that people can understand from experience). Tables, ext links, references, markup, further reading, and HTML entity bloat could be excluded from the size measures since readers get tired by reading scads of prose not the wikicode. I also love your examples and your fresh input - I'll incorporate that into Wikipedia:Summary style later (as I've already done with your 'zoom' statement). :) --mav
The problem with the transclusion idea, as well as with the argument against lenthy summary style articles, is that a good article is greater than the sum of its parts. A summary article presents an opportunity to synthesize the detail contained in detailed individual articles into a cohesive overview of the whole subject. This synthesis is very valuable -- we want to show a picture of the forest and the trees, not just show all the trees standing next to each other. As an obvious example, we want an article on the Olympics, not just a bunch of articles on individual events grouped as "part of the series on the Olympics". Strictly limiting the size of such articles places an undue restriction on this synthesis process, especially when there is a lot of material to cover -- precisely the situation when a good synthesis of material into a summary article is most needed.
Our responsibility as writers and editors is to both the reader and the topic. We cannot write a 200k article and expect it to be accessible to the vast majority of our readership. At the same time, we cannot write a 10k article on World War II that makes any effort at effectively synthesizing even a few of the important aspects of that event. In the end, readers are cheated if we do not synthesize an effective overview of a complicated topic, and the topic is cheated if we make the information inaccessible by creating an unreadably long article. The size of a summary article must be dictated by the subject, not an arbitrary limit. If there must be a limit, let's set it at the longest any of us could imagine would be needed for an adequate synthesis of the most complicated article -- but we would know that that limit is much too high for most topics, and therefore would allow more unneccessarily long articles than it would prevent excessive size.
We are also failing to acknowledge readers' abilities to parse long articles on their own. Frankly, only very simplistic readers will simply stop reading if they get bored without reaching the info that they want. Those people will likely be served adequately witha good lead section alone. Most others will skip sections that they find uninteresting to find what they want to read, or use the TOC, or briefly scan the entire article. And of course even quite long articles don't seem very long when they are well-enough written (not to hold this up as an example of fine writing, but how many people consumed the Harry Potter novels hundreds of pages at a sitting?).
(Sorry for my absence and the multi-pointed post; I was on brief hiatus to lift heavy things for my mother-in-law.) - Bantman 21:32, Apr 4, 2005 (UTC)
Nobody is talking about having articles as small as 10KB or as large as 200KB. What is currently being proposed is a range for an upper limit to article size. This will allow for some variation in the amount of content needed per topic while at the same time making sure articles are at a length that will not exhaust readers. Lead sections are great for those needing a concise article on a subject, but many others need a mid-level of detail ; not too long to read in a single sitting but not too short in terms of content covered. Yet others will need a good deal of detail. All three reader-types need to be served. --mav 15:28, 5 Apr 2005 (UTC)
I was hyperbolizing with the 10k / 200k numbers, didn't mean to add them to the actual negotiations on size. My main point was that good coverage of a topic requires synthesis of detailed information into a cohesive and insightful overview, and that this requirement does not neccessarily fit into the "small / medium / large" scheme for serving different types of readers. The most detailed information should indeed be split into separate articles, but the summary article -- what I suggest should be a synthesis article -- should be as long as it needs to be to serve its purpose. I don't think we can establish what that size is without seeing individual articles.
Setting an upper size limit, or even a range of limits, presents another problem. Right now, we treat all articles under 32k as acceptably sized, and all articles over 32k as inherently suspect. If we were to move the size limit to, say, a range of 45k - 65k, we would make it very easy for any article up to 44k to skip size discussions altogether, allow articles up to 65k with the merest modicum of justification, and disallow articles 66k+ altogether. Thus minor topics can swell to a rather large size, going way past "comprehensive" and into the territory of "exhaustive", without triggering a review of article size; significant topics will grow unneccessarily over time toward the upper limit because anything within the range is considered acceptable; meanwhile, huge topics will struggle to reach the threshold of "comprehensive" while trying desparately to squeeze in under the upper limit.
Therefore I propose that rather than setting a hard upper limit, or range of limits, we add more subjective guidelines and allow editors (and FAC voters) to interpret them on a case by case basis (these are just working ideas):
  • Treat topics concisely: articles should be comprehensive but not exhaustive.
  • An article should not be longer than it needs to be to properly discuss its topic in the appropriate amount of detail given the article's style.
  • Summary articles should synthesize the information contained in detailed sub-articles into a cohesive overview of the subject.

- Bantman 16:56, Apr 5, 2005 (UTC)

I think those are fine guidelines. As I've argued above, we should avoid quantification. I would make one emendation to what Bantman has said, however: I suggest that instead of saying,
  • "Treat topics concisely: articles should be comprehensive but not exhaustive"
we should say,
  • "Featured articles should be concise. Articles that survey a major topic should be comprehensive but not exhaustive."
I suggest this emendation because daughter articles probably can afford to be more-or-less exhaustive and still not be overlong. Hydriotaphia 18:01, Apr 5, 2005 (UTC)
Better than nothing. Using the word concise is going to be a bit confusing though, since a concise encyclopedia article is the same size as a lead section. Instead using the term tightly-focused summary would be better. Hydriotaphia ; each daughter article can be a survey article for its part of the topic. The survey/daughter relationship is relative, not absolute. --mav 15:53, 11 Apr 2005 (UTC)
Concise is the whole point, Mav. A "concise encyclopedia article" has no established definition as a three or four paragraph article as far as I'm aware of, and certainly not that the WP community at large recognizes. I intend to convey the idea that articles should not be any longer than they need to be to explain their subject matter in appropriate detail given the scope of the topic. Nonetheless, I'm open to suggestions on improvement of language -- perhaps "succinct" is better? "Compact?" Personally I prefer concise, but that may just be me. I think that "tightly-focused summary" conveys a different message than the one I was trying to get across by implying guidelines for the content of the article itself: I believe a very broad (opposite of tightly-focused) article can be an FA; it merely needs to be tightly written, not tightly focused, to keep its length reasonable. - Bantman 17:24, Apr 11, 2005 (UTC)
'Succinct' works nicely. --mav 23:34, 11 Apr 2005 (UTC)
See below. Hydriotaphia 02:26, Apr 12, 2005 (UTC)

Introductions

I'd like to open up a discussion about new criteria for introductions to longer articles. Let me quote what Bantman and Fawcett5 had to say about this:

I strongly feel that certain articles dealing with complex events and individuals will sometimes and appropriately be relatively large. This is OK, PROVIDED that there is an appropriate lead section. Long articles require lead sections that let the non-motivated reader get what they want and get out. So the focus should be on demanding really well-done lead sections in articles that are longer.

I agree regarding the importance of a strong lead section. Indeed, the common three-paragraph limitation on lead sections is a problematic restriction on long articles . . . . [T]he lead section should be brief . . . but comprehensive enough to A) properly introduce the article; B) provide a quick summary of the article to those with really short attention spans; and C) allow a reader with a narrow topic of interest to get a good overview, and then use the TOC to find the section he wants to read.

These seem to me good suggestions. Hydriotaphia 01:55, Apr 2, 2005 (UTC)

Anything more than 3 paragraphs and lead sections no longer really serve as concise encyclopedia articles (I have a copy of EB concise and all those articles are the same size or smaller than our current recommended lead section size). Having leads that are too long also pushes the TOCs too far down the page. And just as leads can quickly summarize a topic in way that is comprehensive for their length, a survey article can do the same at a mid-level of detail. A mid-level of detail is the type of thing you'd expect to find in encyclopedias such as Encarta or World Book, while you'd expect the full Encyclopedia Britannica to have articles that could be huge (I've seen some with 500 sections). Different people are looking for different levels of detail (EB, EB Concise and the other two mentioned encyclopedias serve different audiences - we need to serve each of those same audiences and not just focus on one or the other). With summary style this diverse readership is served.
So the answer is not to make leads longer, it is to make long articles more compact and have them cover just the more important information. Articles on sub-topics that are prominently linked from just below section titles in the survey article is where that sub-topic can be explored in more detail (that sub-topic itself could then eventually be divided in the same fashion). This gives readers the ability to zoom to the level of detail they need. --mav 02:18, 2 Apr 2005 (UTC)
Just to clarify, I was not advocating that lead sections should generally become longer than the current 3 paragraph guideline for a long article.. I think that is usually enough. I just think that for a long article we should demand that the leads be better. I like the idea of having a concise EB style article for the lead. But many longer articles just don't have sufficiently developed lead sections, or they don't use their three paragraphs wisely. I'd also like to further clarify my earlier point about summary style. When writing a biography for instance, an individual may typically play a role in a larger-scale event, and that event usually merits a subarticle...but the role of the biography is to describe those same events through the lens of that individual. To advocate having THREE articles in order to slavishly stick with summary style seems to me perverse (i.e. Biography of person, role of person in event X, and Event X). In other cases, as with The Cantos, it just made sense to put everything in one article, and the TOC does its job nicely. Give the reviewers credit for their sophistication....they are quite able to discriminate on a case-by-case basis. Fawcett5 04:15, 5 Apr 2005 (UTC)

Proposal for size

I think it may be time for a consensus to be built. Those who are just joining this discussion should see the section above entitled "Size." Here, then, is new language about size that I would like to propose should be added to What is a featured article.

  • Featured articles should be concise. Articles that survey a major topic should be comprehensive but not exhaustive.
  • A featured article should be tightly written, that is, not longer than it needs to be to properly discuss its topic in the appropriate amount of detail.
  • If a summary article is felt to be appropriate in dealing with a major topic, it should synthesize the information contained in detailed sub-articles into a cohesive overview of the subject.

Hydriotaphia 02:33, Apr 12, 2005 (UTC)

  • Support this proposal. Obviously. - Bryan is Bantman 16:36, Apr 12, 2005 (UTC)
  • Support Fawcett5 19:41, 12 Apr 2005 (UTC)
  • I disagree with explicit instruction here; this discussion has generated more text than all the discussions I've so far seen on the length of actual FACs. I think it is useless, and the issue should be sorted out on a case-by-case basis. 119 19:43, 12 Apr 2005 (UTC)
  • Neutral. Creating subpages for more detailed topics (i.e. histories of specific battles can be split out from histories of larger conflicts) is only implied. slambo 19:47, Apr 12, 2005 (UTC)
  • Concise can mean something much smaller than I think you intend (the size of lead sections), so is not a good choice of wording since it is confusing. Also, talking about 'summary articles' is very odd since every article should be a summary (almost every FA I've seen can have books written about them). Also the proposed wording seems longer than necessary and needlessly confusing. Adding 'Tightly-written' to point 1 on the project page and have a bullet point that says 'Tightly-written: not overwhelm the reader with more detail than is necessary to properly cover the major points of the topic. Sub-topics can be dealt with in more detail in daughter articles. See Wikipedia:Summary style to see how to deal with this.' Heck, this could even be merged into the 'Comprehensive' point by saying Comprehensive but not exhaustive then have the wording I just proposed. --mav 17:10, 13 Apr 2005 (UTC)

OK - here is what I got so far from the all of the above talk:

    • Comprehensive but not exhaustive: Covers the topic in its entirety without going into more detail than is necessary. So while it does not omit any major facts or details, the article is tightly written, that is, not longer than it needs to be to properly discuss its topic in the appropriate amount of detail. Subtopics can be dealt with in more detail in other articles.

This would replace the current 'Comprehensive' point. I'll replace the current point with the above in a few days. --mav 12:35, 24 Apr 2005 (UTC)

On second thought the above is a bit wordy. It might be best to keep things as they are and just add this point:

    • Tightly-written: That is, not longer than it needs to be to properly discuss its topic in the appropriate amount of detail. Subtopics can be dealt with in more detail in other articles.

-- mav 12:45, 24 Apr 2005 (UTC)

  • One problem I see with this adjustment: you need to define "tightly written" in a visual sense; without a visual aid people are going to interpret "tightly written" in a multitude of different ways. TomStar81 08:40, 30 Apr 2005 (UTC)
    • Not sure what you mean by 'visual sense'. Please explain. :) --mav 16:52, 1 May 2005 (UTC)

OK - back to my original idea. Replace the 'Comprehensive' point with this:

    • Comprehensive but not exhaustive: Covers the topic in its entirety without going into more detail than is necessary. Subtopics can be dealt with in more detail in other articles.

-- mav 16:58, 1 May 2005 (UTC)

  • Let me give an example: To explain to a newcomer how to put images into articles I could:
  1. Waste about three quaters of the page explain how to use the "[" and "]" keys, how to size images, how to incoperate words in the picture, etc OR:
  2. tell them to simply click here for more information on placing images in an article.

The latter of the two examples is designed specifically to show people how to place the images into an article, and gives examples for each of the options. Similarly, what I suggest is that a group of you get togather and write a page that shows line for line what you want the length to be. Doing this gives everyone a general idea of the acceptable size for the article, so that everyone can interpret "Comprehensive but not exhaustive" is the same way. TomStar81 23:46, 1 May 2005 (UTC)

That is going to vary way too much from article to article to be useful. But Wikipedia:Summary style does exist. With some modification that page could be linked to. --mav 16:47, 10 May 2005 (UTC)
That works. TomStar81 21:15, 13 May 2005 (UTC)

Here's another go:

  • Length appropriate - Stays tightly-focused on the main topic without going into more detail than is necessary to fulfill the other FAC criteria. Sub topics can be covered in more detail in other articles (see Wikipedia:Summary style).

I still need to modify Wikipedia:Summary style per the above long discussion on this topic. --mav 04:08, 17 May 2005 (UTC)

Summary style updated. Will make the new criteria go live in 24 hours. --mav 01:01, 23 May 2005 (UTC)

Going live. mav 02:48, 24 May 2005 (UTC)