_Primary review body_
The Primary review body has two purposes,
The first is to reject articles. It will be allowed to reject an article if the article,
The Primary review body's secondary task is to clasify an articles content, examples of possible classification include,
The classification will be stored with the article.
In the long term the Primary review body may need to be split into two part, one dealing with rejection and the other with classification.
The review body will work democratically, each article will be presented to the body and they will vote on it. The review body should be able to grow in size, to permit this new members should be allowed to join the body by a majority vote of the body. It should also be permittable to dismiss members of the body by a vote.
In a vote to dismiss a member, only those who were members prior to that members entry should be allowed to vote, this will prevent possible abuse of the system if a subversive group tried to gain a majority in the review body.
A proxy group could reject/accept articles on whatever criteria they wished, any activities that happen between the author and the proxy group are out of the scope of this document.
These would allow searching of the article database. A possible search engine could allow users to vote if the users thought the article factually incorrect, this data could be stored and a tally of such votes against an article displayed along side the article.
This would be the main way for normal users to access the articles. Anyone should be allowed to create an Index. An Index would basically be a collection of article-IDs pre-selected by the people running that Index. This would allow a form a moderation without causing censorship.
A possible style of indexing could be in dmoz style where the Index is controlled by one person (the "primary controller"). This person would from a top-directory level create subsections. The primary controller could either manage these subsections themselves, or appoint controllers for these specific subsections. The controllers of these subsections will have all the ability of the primary controller except that their powers will only apply to their subsection.
The controllers would be responsible for adding links to articles to their section of the index and also for creating subsections and appointing subsection controllers.
A controller would have the ability to alter everything in their section of the index, including anything in subsections of their section.
This will allow a tree like structure of trust with the ultimate authority with the person who started the index.
There could exist many of these indexes each favouring different types of articles and different styles of viewpoint. For instance an index could be setup which only dealt with non-controversial areas.
If Nupedia wished to integrate with this system it would be relatively easy. Nupedia could become a "proxy group" with Nupedia accepting articles as they currently do and then passing them into the system. Nupedia would also run their own index which would be primarily based on articles signed(as proxy) by Nupedia.
Nupedia in this way could become a factually reliable front for the encyclopedia which provides a strong moderation style. Nupedia could also have people searching through the articles to find articles which would pass Nupedia's style of filtering, Nupedia could contact the authors of these articles and request the author resubmit the article via the Nupedia proxy group. Most authors would be likely to accept such an offer as Nupedia reputation as a reliable factual source would make it more likely that the article will be accepted by other indexes.
I think this would provide the best combination of the GNU Encyclopedia/Nupedia? as it would emphasis the strengths of both without comprimising either.
Comments appreciated, preferably via the bug-gnupedia mailing list, but I can also be contacted directly at ImranGemail@example.com. Or list below:
I added Imran's article above. I'm uncertain about the classification approach ("mainstream" seems hard to define, to me). I would advocate simply advertising a list of preferred classification tags, and let the author decide for herself where they'd like to be classified (if at all). Look at www.indymedia.org for an example of a system where they must balance the need to be open to all, yet filter out blather. Their approach is, rather than vote on each item, to use a traditional editor, but keep the process open (anyone can look behind the scenes to see what was not accepted.)
Also, as a *great* example of how indexes can be constructed dynamically and collaboratively, look at the front page of THIS site. :-) The Nupedia folks laid out their idea of how to hierarchically index the material here, but anyone coming to this page can generate their own unique index. E.g., I could say I want my encyclopedia to have this set of topics be my index: