Wikipedia talk:Root page

Roots vs. dabs

I like the idea of root pages. Ambiguous names rarely come about by accident, and we ought to tell the story of how they originate. At the same time, we need good old disambiguation. The encyclopedia would just break down without them. We need a good way to combine root pages and dabs, without the one competing with the other. --Smack (talk) 06:45, 6 March 2006 (UTC)

No, they are totally different. The Root page concept is about ordering pages according to semantics(which topic 'includes' which) for easy navigation and coordinated editing, using a special linking method. Disambiguation is for totally different meanings of the same word. This point has come up again and again. Please have a go at navigating around Electronics or Noise and I think you will grasp what this is about. --Lindosland 11:33, 6 March 2006 (UTC)

It looks great while you have only one root page and one person using the schema. But what will happen when computing people try to include analog multiplier into their hierarchy, the music people try to link audio filter in theirs, and the career counseling guys try to put electronics as a career in theirs? Or when the image processing people get around to edit the article on noise in TV images?

Once more, the universe is not a tree, nor is knowledge. A tree-like picture of either would be misleading and confining. And I still don't see how this scheme can possibly work in the Wikipedia context. All the best, Jorge Stolfi 01:45, 7 March 2006 (UTC)

Could I derect interested parties to the main discussion on this subject at the talk page for Wikipedia:WikiProject Electronics. THanks.--Light current 01:49, 7 March 2006 (UTC)

This proposal will affect the whole of wikipedia, not just Wikipedia:WikiProject Electronics; and the main objections (such as clash between different subjects) are more likely to come from outside. So methinks that the discussion must be carried out here, not there. Jorge Stolfi 23:54, 7 March 2006 (UTC)

This discussion is no longer just about Lindoslands suggestion but about how large articles might be organised. Since Electronics was the most appropriate guinea pig in this experiment, it is only fitting that discussion is carried out there. Otherwise, there will be two discussions needing twice the amount of user input, or lack of communucation. Neither of these is desirable.--Light current 00:24, 8 March 2006 (UTC)

Yes, Jorge, I have realised these difficulties, but still feel there are bigadvantages in 'attaching' pages to a root. The answer to 'what happens when computer people try to link analog multiplier into their area is that they get a jolt - it is already taken! This causes them to look up Root page and get some advice. They might then, I suggest, make their own page multiplier (computing) or whatever with the slant on analog computers, linked via 'see also' to Analog multiplier in Electronics. This is what happens already in many cases, but often without each party even realising that the other article exists. Thus I have attached Optoelectronics to Electronics, but would leave Photoelectrics elsewhere, with its more fundamental slant. I would even suggest that articles might be created to 'complete the set' around a Root page, even if they just contained lists of 'long range' links to other fields. Like LightCurrent I constantly feel the need for a method of easy navigation. Its not hierarchy, by the way. As someone pointed out, its semantics - meronymy and holonymy, a matter of 'inclusivity'! --Lindosland 12:13, 7 March 2006 (UTC)

Dear Lindosland, they get a jolt, it's taken is not an answer, it is THE problem. In Wikipedia, a person (or a group, like the editors of WP:WPE) cannot assume that he "owns" an article just because he was the first to "claim" it. (One could say that it's actually the opposite: the "owner" is the person who made the last edit, and his "ownership right" expires with the next edit.) I would say that your problem is not lack of means, but excess of desires. When you wish for "an easy way to navigate around the electronics pages", you have already assumed that (1) "electronics" is a well-defined category (every article either is in it, or not in it), and (2) there are readers for whom navigating within that category is more interesting than navigating out of it. These sttements may seem obvious to you while you are immersed in WP:WPE, but they are quite false in general. First, there is no sharp boundary between electronics and non-electronics. Second, each reader has a differnt area of interest, and looks at the subject in a different perspective. Third, no one reads 50 (or even 5) encyclopedia articles on the same topic, one after the other, as if he were reading a novel. Wikipedia is not a collection of textbooks: if you folks at WP:WPE are trying to create a textbook on electronics, you are doing it in the wrong place. More generally, I am alarmed by the trend towards compartimentalization in Wikipedia (of which this proposal is only one facet). While the principle "articles have no owners" seems to be generally understood and respected for individual editors, we seem to be moving towards a situation where certain communities feel that they "own" a certain set of articles, like "electronics", "Brazil", "History of the penny", etc.. The first symptom or cause of this problem is the use of special (and usually exclusive) linking mechanisms to connect all the articles of the set. That is not only useless (for the above reasons) but actually deleterious.

I am coming to believe that navboxes of any kind — even trivial ones like those of History of Brazil or the {{main}} templates — are the devil incarnate. In fact, I dare say that if all the navboxes in Wikipedia were deleted overnight, no one would miss them (except the "group owners", of course) and Wikipedia would suddenly be ten times better.

The root page proposal, specifically, besides establishing a group, also creates a tree structure within that group. Let me say it again, more emphatically: no matter what the subject, a tree structure is guaranteed to be the wrong structure. (The only exception may be biological species, where the tree is wrong only some of the time; but genetic engineers are already working to "fix" that 8-).

The limitation of tree structures is well known to those of us who still use 20th century operating systems, like Linux or Windows, and their tree-like file systems. In those systems, for instance, one could not store email messages grouped both by sender and by topic: it was either one or the other (and each message could have only one topic). Thus files were always in the wrong directory, and we often had to create two or more copies of the same file just to get around that limitation. (Linux had a thing called symbolic links that solved some of those problems, and created a many more.) Lucky of you who have modern file systems...

Speaking of making copies, the suggestion that the coputer guys create a copy of analog multiplier is an absolute no-no.

By the way, I have heard that there is a new navigation tool in Wikipedia called "wikilink". It is said to be extremely powerful: you can use it anywhere in an article, even in the middle of the text, to make any phrase point to any other article. Its inventors claim that wikilinks can be easily combined to make bulleted lists of links, trees of links, and even two-dimensional arrays of links. Sounds too good to be true...

The main disadvantage of this tool, it seems, is that it occupies zero screen space, so it is nowhere as eye-catching as a navbox. It seems that only those readers who actually look at a paragraph, bullet list, or figure caption will notice the wikilinks contained therein; thus editors may be forced to put wikilinks like perverb in those parts of the article which are likely to be read by readers who may be interested in perverbs. Another disadvantage is that the removal of a navbox leaves a conspicuous empty space at the top right corner of an article, with is likely to get vandalized by the insertion of a nice picture or map. (Some editors have also complained that wikilinks tend to allow the reader to jump to articles he is interested in, instead of those that the editors want him to read.)

Nevertheless, in spite of those problems, I would say that wikilinks are a great improvement over all their successors. I wonder whether they will ever become popular with editors.

All the best, Jorge Stolfi 23:54, 7 March 2006 (UTC)

Dear Jorg Stolfil: I understand your point of view, and support it to some extent, but I do not think you quite understand mine, and you mis-represent what I have said. I did not, for example, propose making a copy of Analog multiplier, I proposed a solution involving a pair of articles with different 'slants' as I find is often already done.

I have said over and over again that Encyclopedic articles are not hierarchical in any one dimension, but they do have semantic relationships and there is no doubting that many fields like Electronics 'include' (Holonymy) many topics. Like it or not we live in a world of compartments, and I wonder if you think Universities could function better if they had no departments and no courses like electronics, biology, etc and students just chose their mix. The problem of course is that without compartments you have no lab facilities, no close interactions between like minds, no peer review.

That said, you should know that I am deeply involved in Genetics and Evolutionary theory, and a keen supporter of E. O. Wilson who proposes the idea of Consilience (I went to his lecture on it). I also understand the problems of biological classification (cladism), where a similar problem resists easy solution. I hate the way people are categorised as engineer, biologist etc, but I do not oppose them working together in groups under such a heading. What all this leads me to conclude is that compartmentalisation suits control, keeping order, and team work. Lack of these things opens things up, but with the danger of anarchy. As with everything in life, we seem to need a balance of the two.

No, of course Wikipedia articles are not 'owned', but in practice things have, as you commented, tended to go that way with groups watching over articles. I suggest that it is this that makes Wikipedia work as well as it is doing. Is it not the case that the freedom you support clashes with the NPOV rule, and it's the 'peer review' operating on pages that 'forces' NPOV?

I confess this is not simple, but I think you should give more weight to the editing problems. Coordinated editing produces good pages. You can see that all over Wikipedia, and you can also see disjointed articles with wrong ideas - and they are usually the ones with a blank talk page. On the user side you say that no one reads through fifty articles on one topic, but take away the exageration and I suggest that they do indeed read several. I gave you a good example in Semantics. Try sorting out in your head the subtle meanings of homonymy, synonymy, antonymy, hypernymy, meronymy and holonymy (a list given on this page by a contributor) and then tell me that a branchlist such as I would put on each of those pages would not be a huge help! It also occurs to me that I have looked through many many pages of Genetics where experts have put up so much detailed stuff that I Wikipedia starts to take on a new role - as a teaching tool that is bang up to date. You would need access to a library full of journals to get at some of the information there, as it has yet to get into the textbooks, and this is going to be the case with many other fields too, including Electronics. Wikipedia doesn't have to stay like other encyclopedias, I reckon its soon going to be 'the one'. I read that many people already use it constantly in their work. I don't want to push these ideas through if they are not right, and I am beginning to wish I'd never proposed them, as they turned out to involve a lot more thought than I had expected, but the more I look into it the more I feel that I've raised some very relevant issues, which it is good to have discussed here. Regards --Lindosland 11:48, 8 March 2006 (UTC)

By the way, I agree with you that discussion should be kept here. Electronics is by no means the most relevant article to demonstrate on, there are thousands, and yes, this does affect the whole of Wikipedia. --Lindosland 11:55, 8 March 2006 (UTC)

Regarding {{backlink}}

I was looking up some information regarding Motion capture today, and I happened upon the {{backlink}} template. I didn't know anything about this "Root page" project, so I hope that my opinion will be valuable as an outside view.

My first thought when I visited the page was "What the heck?" The template made no sense to me. I had gotten to the Motion capture page from Google, so seeing the words "Back to..." at the top was somewhat jarring. I was not interested in general Animation topics, and if I were, I would've just searched Wikipedia for them.

I feel very strongly that this template disrupts the overall Wiki structure and experience; it is counterintuitive based on the way I browse. I believe myself and many others do not browse in a linear fashion, rather, I skip from one article to another, sometimes searching, sometimes using bookmarks, sometimes clicking links. Saying "back to" only makes sense when you've arrived at an article from the root page, something I would speculate as being a rare occurrence due to the above statement.

I don't think that having a link back to the root article is entirely out of place, but I also don't think it should be placed at the top of the article. I believe the most appropriate place to put the link would be in the See also section. ~MDD4696 01:11, 14 March 2006 (UTC)

Copying a very relevant post by The Photon from Wikipedia talk:WikiProject Electronics#Root page proposal:

Tools are already available :

We already have several good tools to use to tie a group of articles together.

• The first is just good writing, and the use of explicit crossreferences.

For example, the Diode article might begin, "In Electronics, a diode is a...". This achieves everything that a backlink template does, without distracting from the article in front of the reader.

• Another tool is Wikipedia:Summary style.

Since it doesn't create categories, this isn't exactly the same for editors. But for readers, the experience is the same. Whether the main article in Summary style must summarize the detail articles is just a matter of style. If there are really hundreds of detail articles (as in Electronics), the summary could naturally be reduced to a single sentence, or membership in a list.

• A third is the Wikipedia:Navigational templates.

A well-designed template could tie a group of articles together for its editors, even create a category for those articles; and it wouldn't distract readers.

• A fourth tool is list articles.

A central hub list could be created to allow editors to keep track of all the articles relevant to their area. If it is really just for editors, it could go in the User Talk namespace.

• A fifth tool is Project pages.

Use the Wikipedia:WikiProject Electronics page as a hub for all the electronics articles.

• A sixth tool is "What links here?".

If you, as an editor, want an overview of the articles related to electronics, just click on "what links here".

• A seventh tool is the Wikipedia:Portal page.

Articles are the main organizing structure of Wikipedia:One of the most enjoyable things about reading Wikipedia is the interconnectedness of the articles. By following links, you can explore a dense mesh of related articles.

The proper place for a larger document, that leads a reader along a particular path to understanding a broad field, is Wikibooks. Good writing is the best solution for readers: If the article is well-written and well-linked, it should be clear without any props where to go next for more information about a broad topic. So if in doubt, just write the articles clearly.

The Photon 03:35, 1 March 2006 (UTC)

~MDD4696 01:33, 14 March 2006 (UTC)

I want this template deleted. The demon of hierarchy, which had been staved off when Subpages were deprecated, once again rears its ugly head. --Smack (talk) 03:21, 21 March 2006 (UTC)

Which template exactly, and where? All the templates I put up were valid as navigation templates under Wiki policy. By chosing to list only a selection of all possible pages, according to principles stated here I did not prevent anyone from doing a full listing, as is done already on many pages, I simply offered a neater version. Do you want all navigation templates deleted? --Lindosland 16:34, 21 March 2006 (UTC)

He was probably referencing {{backlink}}. Smack, the template was nominated for deletion and consensus was for deletion (see its talk page), but it was kept for discussion here. ~MDD4696 23:23, 22 March 2006 (UTC)

Shortcut

I have added the "shortcut text" to the page: WP:ROOT. Somethig simple, but normally needed; I hope this will make going to this page quicker and easier. Kilo-Lima|^(talk) 17:04, 18 March 2006 (UTC)

Root pages and subpages

You all realize that Wikipedia used to have root → sub functionality, right? It was intentionally removed. Read through Wikipedia:Subpages and all of the links at the bottom. — Omegatron 00:38, 22 March 2006 (UTC)

No I didn't realise this. Nevertheless, I remain convinced that the concept adds to Wikipedia's functionality greatly, while TAKING AWAY NOTHING. I note that the function was ended because of the problem that sub-pages are often perceived as being related to many other pages, and this is a problem we are well aware of, but I have attempted to solve it by stipulating that a page can only 'belong' to one root page. This does not seem to have been considered before. While some people seem to feel agrieved at the idea of a page 'belonging' to a root page, and suggest that editors may be competing for 'possession', I still do not think this need be a big issue in practice, and the idea that we are taking away 'freedom' in some way is mistaken. I think it is better for a page to be attached in a grouping of some sort than none at all, provided we are all sensible about it and realise that there must remain a degree of arbitrary decision in the matter. I see that most of the nav templates I put on Electronics have been left in place, and I find them so useful that it would seem very retrograde to take them down. Yes, some pages could equally belong to other topics, but it seems clear to me that we lose something and gain nothing by taking away the templates. Most of the Electronics pages were already in the Electronics category and no other, suggesting no conflict. The few, like Electronic noise that fall across two topics, are easily resolved by the creation of two pages, with different emphasis, each pointing to the other in 'see also'. In fact, such topis tend already to have two pages, as one of the main weaknesses of Wikipedia is failure of editors to realise that their topic actually exists in a different form in another field. Thus, when I first came to 'nIn the extreme we could solve the problem with 'shortcut' pages - pages that just contained a redirect to be used much like the 'alias' or shortcut on our computer desktops. I have not the slightest desire to tie Wikipedia down to an imaginary hierarchy, but it is clear to me that the benefits of grouping pages in the way we think 'most obvious' are enormous. The potential for the automatic generation of the nav templates is also very attractive. The problem of chosing the 'most obvious' grouping is very similar to the problem of choosing the most obvious topic when considering disambiguation. The official advice there is 'choose the topic the user would expect to come up', and that's what I advise for grouping pages - chose the root page the user would most expect the page to be attached to. In both cases there will be some people who think differently, but what does it matter how a page is grouped? --Lindosland 19:47, 22 March 2006 (UTC)

What evidence do you have that "one of the main weaknesses of Wikipedia is failure of editors to realise that their topic actually exists in a different form in another field"? Once someone does realize the connection, why isn't adding a normal link in the See also section or merging the pages sufficient? And to your point about choosing the most obvious groupings: Categories already allow us to group pages, and put them in more than one group without conflict. ~MDD4696 23:22, 22 March 2006 (UTC)

• The evidence of my own considerable experience in editing pages. POV is allowed here! There is a very good reason why adding a link as you suggest isn't sufficient. I find over and over again the same thing cropping up - pages written by people who limit them in their definition because they do not realise that others use the same word differently. They conflict. With more centralised editing they would explain and expand. For example, until today (when I re-wrote its introduction) Noise (acoustic) defined acoustic noise as noise pollution. I say that acoustic noise includes noise pollution, along with many other forms. Again, until I changed it, White noise would have us believe that this was a purely electronic pheonomenon, but I can assure you that audio engineers use acoustic white noise in testing. Pages that take a myopic view that conflicts with other pages need modifying. Its a sort of Wikipedia syndrome (Myopedia?), and the irony is that in the long run it's Wikipedia that could cure it, by making editors aware of the bigger picture, and then users. --Lindosland 23:42, 27 March 2006 (UTC)

Actually, the subpage system only allowed a page to be a part of a single parent page as well, so your system isn't really any different than it was. I don't actually see any major differences between subpages and your system, aside from the navboxes, which don't fundamentally change the idea. - Flooey 06:19, 24 March 2006 (UTC)

I think this whole root/branch structure is terrible. I won't bother stating my reasons, which have been well explained by many others here and in the archives. I would like to comment on Lindosland's repeated suggestion that articles falling into two "root" hierarchies are best handled by forking the article into two separate articles (see his comments on Electronic noise, above). This is really stupid, and the very fact that he suggests it is one of the strongest arguments for dumping the whole root/branch thing as fast as possible. One of the strengths of Wikipedia is that it builds connections between fields. An article that is relevant to editors in two fields will eventually reflect both points of view, and can then be a benefit to people in both fields and encourage "browsing" between the two.--Srleffler 04:09, 23 March 2006 (UTC)

• How strange that you call the idea of 'forking' really stupid, and then seem to go on to say that this is what will happen anyway and that this will encourage browsing between the two. I find that what you call forking all too often exists as two articles which contradict or mutually exclude each other's definition of a topic. See my above comments on 'Myopedia' or short-sighted editing by those who think a topic has only the meaning that they happen to know. --Lindosland 23:47, 27 March 2006 (UTC)

You seem to have misread me somehow. Nowhere do I say that forking will happen anyway. What should happen instead is that when the same concept is used in several fields a single article should develop that reflects both/all points of view and compares and contrasts them. This emphasizes the similarities between fields and encourages and supports "browsing" from one to another. The fact that you are encouraging separating a single concept into separate articles, each reflecting one field's view of that concept, is for me a strong reason to trash this entire "root page" approach. It does not suit the way information is best organized on wikipedia. Your root page idea encourages myopic editing, by attempting to put pages into subject-area boxes. --Srleffler 00:36, 7 April 2006 (UTC)

With all respect to Lindosland, I agree with those who think this is a bad idea. In particular, I think there is a lot of POV context in deciding what article precedes what article in a hierarchy. For example, Ophthalmology was previously edited to backlink to Optics. In my opinion, optics is only a part of ophthalmology and should not precede it in a hierarchy. I do think Srleffler makes an excellent point. I've noticed that he and I are in different fields but now and then share some overlap in the articles we monitor. I would like to believe that Wiki benefits from the different take each of us brings to those articles. -AED 07:05, 24 March 2006 (UTC)

What is "linking"?

There are lots of references to "linking", "backlinking" etc. What does all this mean? Is this talking about infoboxes, category links, normal wiki links, or what? Stevage 09:01, 24 March 2006 (UTC)

The term backlink got introduced by others, as this idea evolved. Otherwise, I think the project page expains what is meant quite clearly. No, this is not about anything else, it's about something new, though I confess that attempts to please objectors have led us round in circles somewhat, and the 'backlink' idea has now been dropped in favour of just a special form of the Navigation templates. Just take a look at Electronics to see what is now proposed in action. --Lindosland 13:00, 27 March 2006 (UTC)

Why this is a bad idea...

Unfortunately, although well-intentioned, this project is not a good idea. Here's why: information is not inherently hierarchical. Different topics cross-cut one another, and it is often unclear whether one topic is a sub-topic of another, or vice versa, or whether the two are actually pairs. The semantic relationships between articles do not form a tree of succesive specializations, but have a vastly more complex toplogical structure, with multiple types of semantic relationship existing between different sets of objects, including asymmetric, symmetrical and cyclic relationships. See ontology (computer science) for why this is a hard problem.

As other posters here have stated, Wikipedia used to have a page/subpage structure, and, after extensive discussion, it was removed for exactly these reasons.

According to Wikipedia:Subpages:

"Subpages were originally used on Wikipedia to differentiate between subjects to create topical hierarchies of articles, but this proved unworkable because articles tend to belong in more than one hierarchy. The present system of disambiguation was adopted instead, and the Wikipedia:Do not use subpages policy had to be rigorously enforced, as well as retroactively applied. The new (mid-2004) category system supports hierarchical organization while still allowing an article to belong to multiple categories."

Since that time, both categories and portals have been added to Wikipedia, and already seem to accomplish much of what you are seeking to do: perhaps you might be interested in seeing how you could best work within these frameworks, perhaps as a WikiProject? -- The Anome 11:09, 24 March 2006 (UTC)

It took us a while to sort out the sub-page mess last time. Please don't waste all that good work by re-introducing them under a new name. Your imagination may tell you that they are a good idea. Our experience tells us otherwise. -- Derek Ross | Talk 16:03, 24 March 2006 (UTC)

• Portals, like big Navigational templates, seem to clutter the page terribly, and I am surprised that they should be considered preferrable. Anome: you do not say why it is a bad idea - I have repeatedly stated that Wikipedia is not hierarchical, but that does not mean necessarily that a simplified system of grouping and navigating is not useful. What does it take away? I have now re-written the topic page for Root page as the many variants were becoming confusing. Please see below.

Major Revision in the light of opposition

I have revised the topic page in an attempt to reduce confusion and make the concept more acceptable to those who appose it.

I acknowledge the validity of many of the objections above. In particular:

• The 'backlink' at the top of the pages, originally proposed by someone else here, is now abandoned as it was not liked. The 'This is a Rootpage' template is also abandoned.

• Attempting to 'backlink' for the sake of it, eg from Electronics to Electricity is now abandoned as being annoying and confusing without adding functionality. The Root page concept is now strictly three-level, with Root, Hub, and Branch pages to clarify discussions (but note that a Hub page is still a 'branch' of a Root page in the relational sense).

• The laudable desire of some editors to retain the independance of pages is acknowledged, with a recommendation that the Root page concept only be used where it seems obviously warranted.

I would like to stress that as it now stands, the Root page concept only makes its presence felt through the appearance of a Wikipedia:Navigational template that is much smaller than usual, with up to two topics in the first line (Root and Hub). I suggest that this can hardly be regarded as in any way breeching Wikipedia policy. For those who dislike nav templates it makes them more acceptable by reducing their size. For those who want an easier way of getting an overview of large topics it offers simple navigation without any added text (apart from the nav template). I suggest that this concept should remain, in its current unobtrusive form, for those who wish to use it, and that this talk page constitutes a very relevant bit of Wikipedia history on a very tricky and controversial idea of hierarchy and relatedness among encyclopedia topics.

This idea sounds much better than it did before. In fact, it's so good that it's already been implemented; we call it Article series. All you're suggesting is to allow nested series. On the face of it, I don't see anything wrong with nested series, but neither do I see a great benefit over "flat" series. --Smack (talk) 22:22, 27 March 2006 (UTC)

No, it's not already been implemented. I find it frustrating that so many people are keen to tell us that the idea already exists, or that it has been tried and rejected, without taking a proper look at the whole concept. Article series makes no mention whatsoever of nesting, and that is not a negligible idea. Nor does it mention the purpose as being to aid centralised editing of big topics, or to aid navigation by users. Nor does it aim at automation of a major part of the linking process throughout Wikipedia! And then this is not 'just' nesting, which implies unlimited linking backwards and forwards, and which as has often been said, has a lot wrong with it! Without the right rules nesting confuses, and leads to conflicting 'claims' on pages by several other pages, so that the idea of a 'base camp' for editors from which to take an overview is a non-starter. It can even turn circular, with two pages claiming to be 'above' each other. Recognising the dangers, we're defining a central point, the Root page, and setting a limit of three levels (enough to cover big topics conveniently with lists just 10 to 20 long). We are saying that a page can only have one 'root' or hub, to avoid the problem of conflicting 'claims'. You say you don't see a benefit over "flat" series. Take a wander around Electronics and then imagine doing the same with 200 pages all listed in one huge template on every page. Then imagine branch templates generated automatically in the future, just as the index is currently. Isn't that a great idea? Still, you do say you like the idea. Perhaps after due consideration you'll consider it an advance on anything we have already. Regards --Lindosland 23:17, 27 March 2006 (UTC)

You're right that the reduced root-page idea has not actually been implemented. Article series and reduced root pages are not strictly the same thing; sorry if I suggested otherwise. However, root pages can be implemented very simply within the existing framework of article series.

You'll probably ask why it's worth our while trying to integrate root pages into an existing feature. The main reason that I can see is simplicity, or minimalism. Any new feature makes the wiki process more difficult to learn and practice efficiently. I've been around Wikipedia for a few years now, so I might expect to be familiar with all of the software's features, but people add new ones as quickly as I can learn them. If this trend continues, the growing panoply of features will soon bewilder would-be editors and readers alike. One powerful, generalized feature is better than two narrowly focused ones. See also feature bloat.

Now, back to your question of how to implement root pages with article series. You raise a whole series of issues, which I'll answer in order.

Article series do not claim the goals of centralizing editing, automating linking, or facilitating navigation.

You're right, the page on Wikipedia:Article series does not mention any of this. However, IMHO it's obvious that series accomplish all of this.

• They certainly do not achieve automated linking. Perhaps you have not realised that I want the branchlist templates to pop up completely automatically on pages just as the index does, using a specially modified form of parametric template. This can be done with appropriate software implementation. --Lindosland 22:00, 29 March 2006 (UTC)

Article series do not allow a single article to be identified as root.

This is just not true. See History of the United States. This page, and many others like it, has been a de facto root page for years. Nobody is going to even think about putting up a rival root page, since a well-established root already exists.

Article series are not nested.

This is true. I admit that root pages represent a modification of article series. However, your example of an "Electronics" root with "Analogue electronics" and "Digital electronics" hubs looks an awful lot like a series with two subseries. In fact, we could call it a series and be done with this discussion — it's just a matter of definition that prevents us from doing so.

--Smack (talk) 02:02, 28 March 2006 (UTC)

Yes, I like this. Using article series, series of series, and so on, to create this kind of structure allows [possibly multiple, intertwining] hierarchical navigation structures to be created (good) without attempting to force a hierarchical structure on the articles themselves (bad). This seems to capture the spirit of your original idea without re-introducing subpages or creating an un-Wiki "single true structure", and it re-uses techniques that are already well accepted. -- The Anome 10:58, 29 March 2006 (UTC)

• This is a lot to reply to, and I feel you are both being too hasty. The point about Root page is that I started it in order to encourage the coordinated editing which I felt Wikipedia was in need of. In the process we worked out some essential problems and solutions in detail which are discussed here for posterity. Only in the final stage of trying to please objectors who disliked the extra Rootpage tag and Branchlist, did I adopt the modified navigational template to get to a neat solution. The Anome talks of hierarchical structures and intertwining, but if he read all the talk here from the start he would find overwhelming objection to such ideas, and many comments to the effect that it's been tried and failed. Only when you put strict rules in place can it work, especially in terms of coordinated editing. Pages have to 'belong' to a topic, if editors are to get an easy overview of the whole topic and see what is duplicated etc. This page is about working out the details, which are more critical than anyone ever realises until they try it out - me included! I learnt a lot by being deliberatly bold (some think a bit too bold) and trying to organise a few big topics. Like the problem of conflicting 'claims' and the possibility of solving this by careful 'forking' to create two similar articles but whith different slants. Like the fact that if we allow more than three levels the concept gets out of hand and the template less neat, while with less than three we are back to huge lists without the advantage of nested listing. And like the fact that if a page is allowed to 'backlink' to more than one Root or Hub (your intertwining) then it all falls to pieces, with pages becoming distant Root pages for themselves! I hope you understand that this idea has been about much more than any existing approaches, and will feel able to support it here, rather than try to graft it on to the existing ideas, losing all the discussion history that is now here for all to see. --Lindosland 21:40, 29 March 2006 (UTC)

By the way, a de-facto Rootpage is fine, and I don't dispute that, but many big topics have lacked such coordination, and by putting the new template scheme into action on some of them, along with a note about it conforming to the Root page concept, I have brought people here, and provided the strict framework and recommendations for implementation by all. History of the United States is also a bit of a special case as its more of a listing of similar pages for different eras than a complicated set of articles like Electronics or Sound recording and reproduction (just organised as a Root page]]. Much less scope for conflict over which Root page each topic should 'belong' to. --Lindosland 21:49, 29 March 2006 (UTC)

Need for care in template placement

Copied from User:lindosland talk, as part of a discussion about not spoiling the appearance of pages:

Fair enough. On pages with no image at the top, I think your navbox works pretty well. When there is an image there, placement requires more judgement, and individuals opinions will vary as to what is best. I have noticed that the Contents box self-adjusts its width, and it can be a good place to put the navbox, at least for screen sizes of 800x600 to 1024x786, which should cover typical screens today except for hand-held displays. The question which will remain is, how do we deal with a proliferation of navigation methods (categories, portals, various navbox templates, "See also" lists, etc.) I guess we will just have to keep experimenting, and probably several methods will coexist. --Blainster 19:27, 29 March 2006 (UTC)

• I agree about careful placement and have today placed several navboxes alongside the contents box just as you suggest, being a little more careful after your prompt! How to deal with a proliferation of methods - replace them with the Rootpage concept! I dislike big navigational boxes/series templates as they clutter the page and are too confusing to be of help. My scheme, with its nested listing, seems much neater and more effective, and now recruits the idea of the navbox and takes it further while attempting to encourage coordinated editing. Hope you will understand my reasoning and support it. Regards --Lindosland 21:19, 29 March 2006 (UTC)

Categories and question

Lindosland said this: "Perhaps you have not realised that I want the branchlist templates to pop up completely automatically on pages just as the index does, using a specially modified form of parametric template. This can be done with appropriate software implementation." - I don't understand what this means? What is this index? Can anyone explain?

• By 'index' I mean 'contents list box' which is already created automatically on pages with sections - sorry I should probably not have called it the index. --Lindosland 11:28, 31 March 2006 (UTC)

Also, how is this whole Root Page idea different from Categories? I thought Categories also formed a non-hierarchal structure with cross-links between different parts of the "tree". In case anyone is interested, I navigate Wikipedia using the category structure using tools like Category Tree and Category Scan. Carcharoth 01:06, 30 March 2006 (UTC)

• I don't want to keep explaining bits of the concept, as I feel its all been gone over many times now. Its about the whole concept, not just 'how does it differ from x' or 'couldn't this be done with y' and you really need to read the topic page carefully and then take a look at the demonstration pages and navigate around them to see how well it works. You also need to realise how it helps coordinated editing, something that only becomes apparent when you actually try pulling together a jumble of pages that constitute a big topic. As the templates are put into place the whole picture becomes clearer and you can instantly check out what is where. Categories are useless for this. Working on Sound recording and reproduction recently, I found that many of the pages there have different categories assigned, from Audio enginneering to Acoustics etc. Root page ties a page into place in a topic. It uses 'hierarchical' listing, in the sense that the lists are nested, but this does not necessarily imply a strict hierarchy in the pages themselves. Talking about it is far less satisfactory than trying it out. It works well. --Lindosland 11:28, 31 March 2006 (UTC)

New format for Branchlist templates

I am implementing a new format for the templates, incorporating the names of the relevant pages rather than just the number and letter scheme I have used so far. I have chosen to use / as a separator, so the format is Template:Branchlist/Root/Hub. An example is Template:Branchlist/Electronics/Analogue electronics. The idea is that in future a template using parameters separated by pipes will be used. There will then be only one actual template called Template:Branchlist and the parameters will determine its contents automatically. Just applying the template in curly brackets on a page will cause the template to appear, using the Root and Branch titles given as parameters, and listing the actual topic page. There are many ways in which this might be done, and I am looking for help from experts in templates and software. One way might be to use a software 'bot' that scans all content regularly and creates templates in the database. Another (better) way would be for the MediaWiki program to check for any Branchlist template when a page is saved after editing, and do the appropriate generation and storing. Generating the template for display 'dynamically' is not possible, as information has to be compiled from other pages. This distinguishes what is required from the existing parametric templates on Wikipedia.

I chose / as the separator after trying the underbar (which comes out as a space and so is not clear). Some form of separator was needed, as page names can have spaces in them. It is of course important that the full page names are used in the template title, as these will be used as wiki links in the navigation box when automation is in place. --Lindosland 13:06, 31 March 2006 (UTC)

Why not just name them as normal navigational templates? That's what they are, isn't it? — Omegatron 14:27, 31 March 2006 (UTC)

No way! As I've explained fully above they are to be generated using parameters, and so need the common part of the name (branchlist) for recognition by the automation software. Then they differ because of the nested form of the listing. Parameters are already used in templates, but not in a way that can be made to do what I describe above, across pages. What is needed now is a bit of software, preferably in MediaWiki itself, which recognises 'Branchlist' in curly brackets, extracts the names between the slashes, and builds or updates the relevant template every time a page is saved after editing. Its a neat and powerful concept. Every time someone specifies a Branchlist template on any page, the template specified by the parameters between the slashes gets updated (or created if it doesn't yet exist). --Lindosland 23:44, 31 March 2006 (UTC)

Nested navigational templates? We already have categories for hierarchies like that. Also, please don't test proposed software changes on live articles. — Omegatron 23:52, 31 March 2006 (UTC)

What is the approved way of testing new ideas like this one on WP?--Light current 00:05, 1 April 2006 (UTC)

Omegatron: I'm surprised by your objections, which suggest you haven't cought on to what the concept achieves, or haven't looked at the trial articles. Categories do nothing like this. I could give detailed examples of where articles lack appropriate categories, or carry different ones from each other making it hard to find associated article. Many article have no links to any other article, and not links from anywhere, so that they exist in a 'backwater'. This suggests that an automated method such as I am proposing here for linking could be a huge advantage. I do not think it is fair to call this 'testing' now, as it involves nothing detrimental in any way, and is simply a way of organising the already approved templates that makes them smaller and more useable. It is also being done after three months of listening to feedback and responding with changes. I have been careful to remove all earlier trials myself. As for 'testing software changes' - how would I do that? I have just put up a proposal and request for help at the village pump, and any changes made will probably be up to the MediaWiki programmers to decide on and implement.

What I suggest is most significant is that there are now around 200 articles with these templates on them, and they have not even provoked any comment at the talk of the pages themselves, apart from a few reversions by one* person. Supporters of the concept here are enthusiastic. Most objectors have given no actual reasons other than comments like 'don't we already have this' or 'I like things as they are' which have been fully addressed. I think it is necessary to take the Wikipedia advice to 'be (a bit) bold' on things like this if they are ever to get off the ground, but I've put a lot of work into considering constructive and valid comments. I took down all added text and moved to the nav box in response to objections that the added text was a nuisance, because this was a valid comment, which I came to agree with. What it comes down to now is this: take a look at Electronics and decide whether you would prefer a (totally legal) nav box with 150 entries, or the current (trial) ones that do the same job with 10 to 15 entries. If you prefer the former, I'd really like to know why. --Lindosland 12:31, 1 April 2006 (UTC)

*I just had a quick look and found reversions by six people. --Heron 20:34, 3 April 2006 (UTC)

I don't see the need for nav boxes either, usually - but it's much more fun to get your name on a pretty template than to just put the links in the text of the article or in the see-also section, like most of the other 6,826,704 -200 pages have. There's never a need for a "navbox" with 150 entries, that's what CATEGORIES are for. --20:53, 3 April 2006 (UTC)

I'd say it's exceedingly unlikely that this feature will be automated in the software anytime soon. First because it can be implemented fairly well with templates, the tools we already have. Second because as an automated addition to the page it would have many of the problems that plague the table of contents, the most important being that it is difficult to make changes to it to suit certain articles. Templates have a lot of flexibility and it is unlikely that the developers would remove that flexibility in order to automate the feature. Christopher Parham (talk) 19:28, 3 April 2006 (UTC)

I could give detailed examples of where articles lack appropriate categories

Add them.

carry different ones from each other making it hard to find associated article

What? The benefit of categories over heirarchical navigation schemes like yours is that categories can be part of many different heirarchies, instead of one rigid arbitrary one. That's the whole reason why subpages were abandoned and replaced with categories.

Many article have no links to any other article,

Add them.

and not links from anywhere, so that they exist in a 'backwater'

Add them.

I have been careful to remove all earlier trials myself.

I had to remove the noise ones after you marked them for speedy deletion and it made the noise articles they were still in look like they were up for speedy deletion.

As for 'testing software changes' - how would I do that?

Create mock-up articles with your idea in them. You could use a subpage of the sandbox, for instance, like Wikipedia:Sandbox/Branchlist examples/Electronics or something shorter.

and any changes made will probably be up to the MediaWiki programmers to decide on and implement.

That won't happen any time soon. Even with things that have unanimous support (and this doesn't), the software changes take forever unless someone who likes the idea programs it themselves. Even then, you'll have to wait for it to be incorporated into the site.

apart from a few reversions by one* person.

I see several people reverting them.

Supporters of the concept here are enthusiastic.

I only see you and Light Current. Who else? From the look of this talk page, I'd say this proposal is very close to a {{rejected}} tag, actually.

take a look at Electronics and decide whether you would prefer a (totally legal) nav box with 150 entries, or the current (trial) ones that do the same job with 10 to 15 entries. If you prefer the former, I'd really like to know why.

I'd prefer the "subpages" to be in the Electronics category or one of its subcategories, instead of an arbitrary list of 10 articles appearing at the top of each page, that only extend to an arbitrary 3 levels deep, which is partially why I removed the branchlist template.

The noise articles are a good example of the purpose of a regular navigational template. There are only a small number of named noise colors, all relatively well-defined. The colors of noise article covers the basics of all of them, but several have their own articles, too, which may someday be expanded to the point that they actually should have articles (maybe a history of when they were "discovered" or whatever). A navigational template would help to tie them all together. What advantage does your branchlist template convey over a regular navigational template in this example? — Omegatron 05:24, 4 April 2006 (UTC)

TfD nomination of Template:Branchlist

Template:Branchlist has been nominated for deletion. You are invited to comment on the discussion at the template's entry on the Templates for Deletion page. Thank you. — Omegatron 16:31, 6 April 2006 (UTC)

To clarify, the proposal at Wikipedia:Templates for deletion#Template:Branchlist is written as a multiple-template nomination for deleting all of the templates in Category:Branchlist. To provide fair warning, and to encourage all interested parties to vote, I have copied the TfD tag to all the articles in this category.--Srleffler 00:39, 7 April 2006 (UTC)

Revised rules - template placement

Many recent objections to the concept have centred on the instrusiveness of the templates when placed at the top of the page, especially where this upsets existing image placement and creates untidy alignments. I am persuaded by these objections, which show how 'live trialling' is useful in revealing unforseen problems. I have therefore re-written the recommendations on this page, under bold headings for added clarity, including the recommendation that the template should be placed alongside the 'see also' listing on the RHS where a blank space normally exists. This is a tidy alternative to the 'Branch pages' listing that was suggested by some, and much easier to update, as edits take effect across all pages of course. I suggest that this nullifies many of the votes for template deletion from people who had not bothered to register this objection here. --Lindosland 13:52, 8 April 2006 (UTC)

No

John Reid 07:46, 9 April 2006 (UTC)

Consensus

This concept has clearly failed to achieve consensus, and the discussion no longer seems to be going anywhere. It seems pretty clear that this proposal has no hope of ever achieving consensus. What is the process for deciding when it is appropriate to apply the {{rejected}} tag?--Srleffler 23:10, 10 April 2006 (UTC)

I think we can add it now. Everyone has rejected it except the two proposers, as evidenced by

• This talk page
  • Its archives
• Wikipedia:Templates_for_deletion/Log/2006_April_6#Template:Branchlist.2FElectronics
• Wikipedia:Templates_for_deletion/Log/2006_April_2#Template:Rootpage,
• Wikipedia:Categories_for_deletion/Log/2006_April_6#Category:Branchlist

A rejected page is any proposal for which consensus support is not present, regardless of whether there's active discussion or not. Making small changes will not change this fact, nor will repetitive arguments. Generally it is wiser to rewrite a rejected proposal from scratch and starting in a different direction.

Note that, if there is an official "process" for adding the rejected tag, step 2 is probably Revert war. — Omegatron 00:26, 11 April 2006 (UTC)