User:OlavN/A boundless encyclopedia

Polypedia is a proposed term for: "An encyclopedia with (potentially) several articles for each entry phrase". An encyclopedia has traditionally one article for each entry phrase, reflecting the smallness and uniform thinking of the author staff. If many independent authors cooperate on writing an encyclopedia, many articles may be needed for each entry - in order to reflect the richness of knowledge and prevent power struggles. Such an article collection may be called a polypedia. This collection must be a global unity, as one index must encompass all accessible article collections.

A boundless encyclopedia

The WWW (World Wide Web) has room for everything. It can't be spammed. Only individual websites open to the public can be spammed, and they are defended by their owners. Writers unable to post on these sites can get their own sites.

Search engines like Google's make it all retrievable. But they index everything in the text - also incidental mentions of the keywords searched for. The best information source on a topic will generally be an encyclopedia article which has a keyword (heading) corresponding to the topic in question.

Wikipedia

The Wikipedia has the largest collection of such articles - millions of them. But it can only have one article on each topic (keyword), so this article must contain a consensus reflecting mainly the conventional view, and alternative views are likely to be expelled.

Article writing is supposed to be a cooperation, like barnraising used to be in some rural societies. But this simply doesn't work on Internet. The barnraising functioned within a closed society, but if done on Internet, somebody on another continent is likely to be a skeptic wanting to tear down the building. The expressed reason may be that it is improperly built, but unfinished buildings are of course improperly built - and an article can easily be restructured in a manner quite impossible for buildings. Wikipedia articles should generally be regarded as unfinished, and they need to be protected by the rule "Don't demolish the house while it's still being built". But this is very often disregarded by the skeptics - who are not required to have any knowledge on the subject.

When an editor has worked hard to write an article, another editor - without any interest or knowledge of the topic - may come and say that the text is too long, and then slash away half of it. He may, for instance, opine that an article on a book should not contain a synopsis of the book. Those interested in the topic will want a synopsis, and those disinterested don't have to read the article.

Another problem is that the subject of a Wikipedia article must be notable - which means prominent when someone negative to the topic attacks an article. Those interested in a subject will need articles on related subjects which are ridiculously obscure in the eyes of the disinterested. No word is too obscure to be indexed by Google.

Readers need Extended Search

The readers need an article search with a much wider scope than a single database of registered articles.

The first scope expansion step is: Let users specify the languages they can read, and the registered Wikipedia articles in these languages should be included in the search.

The second scope expansion step is: Let users specify the websites they want to include in the search. These could be wiki sites, ordinary encyclopedias, as well as other sites with a compatible keyword search system. Wikipedia is likely to be the first of the alternative sites to be included. The WP sites for various languages may be automatically included, based on the languages specified, or the WP sites may be specified explicitly.

A search will then return a list of hits from the included sites. Each hit should give at least a site identification, the article title, as well as the length of the article text. The length is likely to be decisive for choosing what to read.

Writers need protection

The prominence of Wikipedia is problematic: This attracts editors who want to be censors and delete text they dislike. They may try to improve Wikipedia by deleting weak articles, but this reduces the amount of available knowledge.

Wikipedia doesn't protect writers against deleters. Writing a good text is hard work; Deleting it is easy. Wikipedia does have a simple mechanism for reverting a text deletion, but it still takes too much mental energy to confront a deleter who may be using wikilawyering skills instead of knowledge of the article topic. And in article deletion debates, the meatweight decides. They are defacto votings in which knowledge and a serious attitude are not required. A Delete vote may be supported by a funny or emotional comment about the article's first sentence. The second sentence may kill the deleter's argument, but the deleter is not required to read two sentences of the article. Having a funny deletion party is evidently the main thing.

The notability requirement is useful for preventing people from cluttering up Wikipedia with articles on their bands and friends, but when editors disinterested in a science-related topic demand prominence for inclusion, this negativity prevents accumulation of articles needed by readers interested in the topic.

What is needed?

It is not enough to have an extended search of the encyclopedias which in the conventional manner have one article for each keyword. We should also have an article collection permitting many articles for each keyword. Many writers may want to write about the same topic, creating multiple entries under one keyword. The writers, as well as the readers interested in the topic, can now enjoy having articles protected against negative and disinterested editors. There can now be articles in various versions, with original research, essays and other deviations. Quite obscure topics can also be allowed.

There is no need for deleting articles. The Internet has room for everything, and banning articles is a job for the laws of the nations.

If the schools register student's papers in this index, there may be thousands of entries for one keyword, but that's ok. A Google search may give a million hits without causing inconvenience, as the hits are ranked according to importance, and only the most important hits are displayed.

We should consider a fully open collection of articles submitted for inclusion by Wikipedia editors. (No need for establishing another editor register.) The articles may reside in any private website, but if they are to be worked on by two or more editors, they should be stored in one or more multi-user databases like Wikipedia's.

The search mechanism should be more open than a Wikipedia search, but not so open as the messy full-text Google-style search. The articles should have the topic-focusing of encyclopedia articles, and be indexed by their topic-defining headings (and perhaps subheadings) like Wikipedia articles.

Text formatting

If articles are written with wiki software, they will be properly formatted. But also normal HTML web pages should be accepted. There is now a danger of getting annoying background graphics, advertising, animations etc., but this should be handled by the reputation system. And the articles may be read with software which standardizes the text display.

It might be feasible and valuable to include copyrighted articles in the index, or at least links to them, without a permission from the author - but only if the free and copyrighted articles can be kept apart. (No payment may be asked for in the lookup software.)

Cooperative writing

For articles marked as free, text submitted may be freely copied and used by anybody. This is pointed out to everybody doing editing in Wikipedia. An editor may then create and submit an alternative article by copying and modifying (forking) an existing article. The additions can of course be copied back to the original article. Such activities will hopefully lead to cooperation on a common version.

The persons behind the articles should not be called authors, but editors. This will give a seamless transition to cooperative writing.

Each article should (like in Wikipedia) have a discussion page. Editors and readers may there contribute to the article by proposing new text or text modifications. An editor may find that one or more other editors have compatible views on the article topic, and may then invite them to edit the original article directly - and to later revoke the edit rights if conflicts occur.

The writers should be encouraged to use their real names, but using the Wikipedia's editor register - with many anonymous entries - may lead to much anonymity.

A Reputation Rank system

As there may be many articles on one topic, a search result should be a ranked list of the best hits. A quality scoring of the articles will be needed for ranking the articles. This scoring should be done by editors submitting articles within the same field of knowledge. This should render spammers and incompetent writers rather invisible.

The field of knowledge could be defined by letting the editors position the articles within the Dewey classification. Such a peer-evaluation could also include a score indicating how encyclopedic or essayistic the article is, and perhaps other variables like the amount of funniness or advertising. Links to relevant pages should not be regarded as advertising, but ads on the page - particularly the irrelevant ones - will (like inclusion of copyrighted material) destroy the reputation score and push the article out of most search lists.

Such a scoring system will discourage submission of clearly unfinished articles. No matter where the article text is stored, the central database site should contain discussion pages for all the registered articles. Other editors may there suggest text expansions and improvements, and the editors deemed constructive by the main editor may be given permission to edit the text directly. The article should then be stored on the database site or another sharable site.

If a constructive cooperation with Wikipedia can be obtained, WP user pages may be used for this purpose. (These articles will of course not seem to be WP articles.) Illustrations can be taken from Wikimedia Commons, but may have to be cached on the site containing the article.

Topic classification

The editor submitting an article should also submit the Dewey classification number for the topic, as well as an indication of the language used (Wikipedia's language code). This will enable readers to search for a range of related articles (in languages they can read). It will also make the reputation rank more precise, as editors who have submitted articles within related topics will get more weight for their reputation scores. This weight could be proportional to the number of significant Dewey digits shared by the evaluator and the evaluated.

It may be of interest to view advertisements related to the topic. The Dewey number may be used for establishing relationship - in addition to the topic keywords. Broadly defined ads (with few significant Dewey digits) should be easier to filter out.

The search page

Wikipedia searches require simply a keyphrase to be submitted, and the article having this keyphrase as its title is presented. No article with such a title may exist, but a redirection to an equivalent topic may exist. If also this is non-existant, a (Google-style) full-text search is done.

The search page for our new article search should also have fields for some optional parameters indicating

• the Dewey number. A range may be defined by the lowest and highest number in the range, or the first significant digits may be submitted.
• the languages accepted
• whether band names, artist names and fiction work titles should be included or disregarded. Much confusion ensues because names of real world phenomena are also used as names for various pop culture phenomena.
• the oldest and/or newest date for the article content
• the name(s) of filter(s) to be used. These may be made by an organization or authority espousing a certain view: child vs. adult, scientific vs. (religious) belief, conventional vs. alternative... Such a filter may be a list of recommended articles.
• whether (for a commercial product) product information or a (short time) offer is wanted. This enables advertising to be included in the article index - for one day to one year. Advertising should be paid, financing this system (and preventing spamming).

Growth of the system

The system could start as a database containing one record for each article, indexed by the topic keywords. The record should also have a sentence or two of supplementary text for helping the reader to choose between article versions. (A search for a keyword should return one text line for each hit: The hot-linked keyword followed by the beginning of the supplementary text.)

The record should have a field indicating if the article is copyrighted.

There should also be a discussion page for each article.

The system may in the beginning be limited to Wikipedia editors, and later be expanded to get its own editor registration subsystem.

All the articles may initially be hosted on other sites, and the editors will fully control file updates for any cooperative editing.

Articles may gradually become hosted with the system, starting with those having open cooperation. Articles with good reputation may be cached/saved with the system. This system description has to be rather vague, as the properties of the system will depend on the extent of cooperation from Wikipedia.

Could we get a cooperation for creating such a system?