Inter Wiki Charter

If there's ever to be a federation of wikis, it'll need some ground rules. What are they?

What should they be? The rules will depend on the desired outcome.

One set of rules would define a "constitutional" federation, with rules to regulate good conduct. Such a federation may require an elaborate structure similar to modern governments with executive, legislative, and judicial branches. The decision-makers that control key Internet resources seem to be moving in this direction with ICANN.

Another approach is to define a minimal standard for interoperation. At its most minimal, it could simply be a modified form of link, like PyWiki:RecentChanges. Another extreme would be to decide on an XML format which all InterWiki pages would follow. Usenet would be a good example of this kind of federation. There are few central powers ("There is no Cabal."), just widely-available standards and software.

Yet another way is an "open" model, where the content is freely available, and cooperating groups can establish their own federations. These groups could impose whatever rules they desire, but people would be free to copy any content to other systems and establish new rules. This is somewhat similar to the Red Hat (RedHatTheCompany) Linux situation. Several groups have taken the base distribution (freely available and copyable), added or subtracted features, and released the result as a different distribution.

Be careful to fractally apply WikiNature. Perhaps that would mean the federation was either like a giant game of Nomic, or perhaps a giant anarchy. Don't forget that anarchies and communes don't scale because it's difficult for a person to personally care for a stranger except abstractly.

Draft for a minimal "open" federation:

1. All InterWiki sites may copy any information (pages, links, media) from other InterWiki sites, in full or in part. Any site may use any method it chooses to select available InterWiki information. Only technical or resource-related restrictions on copying may be imposed (like rate limiting).

1a. [Legalese defining an "InterWiki site". Probably longer than the rest of the charter.]

2. Copied information may be modified or deleted, with or without notifying the originating site. (Like all InterWiki information, the modified version may be copied back to the original site.) The originating site has no responsibility for the information after it is copied.

3. Contributors to any InterWiki site give all InterWiki sites a non-exclusive non-expiring right to publish their work, and to construct/publish arbitrary derivative works based on their original work. Sites which leave the InterWiki do not retain rights to InterWiki-published material (which must be negotiated with the author).

4. Retaining attribution of "signed" works is encouraged, but not required. Misattribution of works is strongly discouraged, but InterWiki sites are not required to take action. (This is essentially a clarification that arbitrary changes can be made to copies.)

[The original version stops here. Amendments welcome below.]

Contributors: CliffordAdams

It's simpler to use the FreeWebsiteLicense.

I think this proposal is just wrong. As a contributor to C2, I would not like C2 to be part of such a scheme. The principled alternative should be AggregationOnTheClient.

Other sample charters would be welcome. (Perhaps they should start here, and branch into other pages if they become long.)

See also: InterWiki, WikiPortal, StoneSociety, WikiStoneSociety

See also: MeatBall:FuzzyFaq

<SimpleSolution?> : No Government, No Rules, No Central Database or registry mechanism, the method resides within the page as WikiWords: To be continued (a certain period of time lapse is required before continuation - a couple of weeks - additional research and editing required) </SimpleSolution?> for now

Name Collisions:

You would have to deal with name collisions. A string to a VisualBasic Wiki would be far different from a crochet Wiki or a semiotics Wiki.

I have this idea for a ContextSensitiveNameRegistry?. Unlike your DNS system or Wiki itself, where you map one name to one page, you'd map one name to a multitude of different pages. The appropriate page would be determined automatically from context. This would avoid the "Which namespace did I stuff that into?" problem and the name collision problem. Context could be determined by weighted arcs that are adjusted by reward/punishment and back-propagation. Or perhaps through natural language parsing context-determining techniques. -- SunirShah Years later, I'm surprised I was supportive.

Sounds like "magically determine the meaning through methods that haven't been invented yet" to me. -- DanielKnapp, signed for culpability

PlanetMath does this -- BayleShanks

And an army. It needs an army. -- (KenCarpenter, perhaps exhibiting AmericanCulturalAssumption)

An army to implement the RemoteStrangulationProtocol.

Reading that made my day. Thanks. -- DanielKnapp

View edit of August 18, 2005 or FindPage with title or text search