Successful compromise is possible only when the two parties are sufficiently similar-minded. Otherwise, war ensues. If the two parties are of equal strength and stamina, there will be nothing but endless war. Otherwise, the weaker, or less persevering party loses, and the other party wins. That's the nature of things. Some like war for what it is. Others experience war as frustrating and counter-productive. Sometimes there's no easy way around war.

Wikipedia is a war zone. Each article, each section is a battle field of people with opposed agendas, incongruent standards and inconsistent aesthetic perceptions. That's why the Wikipedia experience can seem frustrating and counter-productive. But there is a way around, if what you care about is producing a body of knowledge in a cooperative environment, which is freely accessible to anyone.

In our physical universe there is a single reality: two object cannot occupy the same spot at the same time. In the virtual world of the internet, alternative universes can coexist. The structural foundation of Wikipedia consist of: 1. The Article as the basic unit of structure, 2. The notion of a one-to-one correspondence between an article and a subject, 3. The free, unregulated and egalitarian nature of the editing process.

So there can normally be but a single article per subject. Each article should be consistent: have a single table of contents, a single, consistently formatted bibliography, etc. And editors of all perspectives and stylistic conventions are simultaneously editing the same article. They do not form any hierarchical structure. Conflict between the editors is expected to resolve itself by means of the vague notion of consensus - formal decision making procedures, such as voting, are deprecated.

The second and third points mentioned above have already been the target of serious criticism in the past, followed by attempts to address the problems inherent in them. Wikinfo address the second point, by allowing two or more articles on the same subject, so an editor with a sympathetic point of view towards the subject of the article does not need to edit the same lines as an editor with an antipathetic point of view toards the subject. PlanetMath address the third point by assigning each article a unique owner, who has the final say regarding the article.

I'd like to address the first point (it's really a combined attack on all three points, but the attack on the first point will play a dominant role). Consider substituting the Section for the Article as the fundamental structural unit of an endeavor such as Wikipedia. Instead of writing complete articles, users author single sections, and offer them to the community. Then, by creating a table of contents, linking to sections of choice, a user can combine the sections to unified articles, much like stitching patches to make a quilt, or like compiling an anthology. The user can then offer his/her table of contents to the general public.

A reader can then guarantee that the article he/she reads has particular desirable qualities he/she values, such as being interesting, reliable or consistent with some point-of-view (or lack thereof, i.e. NPOV), by choosing the right compiler. For instance:

• If i hold user A to be an expert in the field of B, then when i'm looking for a reliable article dealing with B-related topics, i'll make sure to read only compilations prepared by A, who surely selected only reliable sections.
• Those who don't wish to be confronted with some content (say explicit portrayal of human nudity) can select compilers who guarantee to have selected only appropriate sections.
• A society of users may draft a system of content policies and style guidelines (think WP:RULES) and tag only those sections/articles conforming to the policies. Different societies may have different policies, or may interpret the same policies differently without conflict between the different societies.

For this solution to be viable, there needs to be a support from underlying software. For instance:

• Everyone should be able to tag a section with a random tag. A tag should be able to vouch for the identity of the tagger; otherwise, a malicious user can ill-tag a section (say, tag an article with explicit sexual images with the tag "no-sexual-imagery"). The system should be able to quickly retrieve all articles tagged with some specified combination of tags.
• A user should be able to search an article, even though the article is a virtual concept, loosely sewn together from dispersed sections. A user should be able to give a search string and indicate a specific table-of-contents, and the system should be able to automatically search for the string in each of the sections mentioned in the ToC.
• It should be possible to associate a single bibliographic listing with many sections.

I believe Wikipedia is a magnificent endeavor, but too primitive and unwieldy to be practical and useful. It will have been an important intermediate stage in the evolution of cooperative encyclopedias. The experience gained from this experiment is priceless, and necessary for the development of better alternatives. Chapeau, and farewell!

Itayb 12:08, 20 April 2007 (UTC)