I don’t think we have a clear shared idea of just what an online encyclopedia within the Encyclosphere would look like.
Mostly, that’s a good thing. We’re not trying to build an encyclopedia, we’re trying to enable a multitude of encyclopedias, and they won’t all have the same form of presentation or the same editorial and contribution practices.
But for the purposes of our planning and discussion, it will be helpful to define some examples of how we think our protocols will be used.
Mode 1, for me, is the “Alternative Wikipedia”. This would be a collection of entries on a very wide range of topics, very open to contributions, many of which would be small edits. Editorial policy would probably be distributed based on some sort of voting. Most readers would access it via a single website, but it could be cloned, and other publishers could either link to its web addresses, or capture specific revisions of articles to link within their own publications, either by embedding them in some collection of their own, or by addressing them using some kind of P2P content-addressable scheme.
Mode 2 would be a “Personal Encyclopedia”. This would be something primarily authored by one person or a small group. They would publish it, either via web or via some P2P mechanism. Contributions from outside could be submitted and would be accepted or rejected by the core maintainers. In case of a significant disagreement, it could be forked and another group could maintain an alternative version of the collection. This is very similar to most open-source software projects. A key aim would be that authors of books, blogs, or other content could link to it as a reference. (This is actually something I want to build).
The third mode might be less identified as a “collection”, but more a way for people to publish ad-hoc pieces of work in a way that is more linkable and shareable than a blog post. I’m not quite sure how that would be consumed, though.
I expect there is a fourth “mode”, and more, though I can’t really think of them. But we want to enable at least these forms.
There are three important patterns of access to these encyclopedias. The first is that you are browsing and searching in a specific collection. The second is that some author, perhaps in a blog, perhaps in a book or a news piece, wants to reference a specific entry. The third is as general search results, from a mainstream search engine or an encyclosphere-enabled search engine not tied to a particular collection.
In the first pattern of access, there would be a “current” revision of every article, determined according to that collection’s editorial policy.
In the second, the author linking an article would either want to link to a specific revision, or, if linking to a collection that he trusted, might prefer to link to the “current revision” of an article in a specific collection. Those are two different kinds of links.
The third case, of a search engine, would probably link a specific revision.
My assumption is that a link to an encyclosphere article should never be just a plain http URI. It should always carry at least an encyclosphere article revision ID, even if only as metadata. That way if the article being linked to is no longer present at the http URI, or has been modified, then there is a way to search it in some other repository (either an archive or some P2P platform). An encyclosphere-enabled reader (which might be a client application, a web service, or a browser extension) would be able to retrieve the revision being linked to, and also indicate that there is a more current revision, or a higher-rated one, or something like that. It would also be able to validate via signatures and hashes that the content linked is what was intended by the author of the source of the link.