Inter-personal relationships are a big factor in the development of a wiki. They are created in a very subliminal way, and they are (as well as their impact on content) hardly perceivable to the casual reader. Therefore, the Mediawiki software should support arbitrarily definable relations between registered users. They might be predicated with a value (like a weight) or more complex forms of data. Some of these relations might be visualized on a person's profile page, others could be used in an advanced visualization of an article's revision graph.
This implies a significant overhaul of users' profile pages. For example, relations could specify banners, stylesheets, or templates that people can use on their profile. Relations could also enforce certain elements on somebody's profile page, e.g. to show that this person works with the government or at a commercial institution.
The basic idea is related to the XHTML Frieds Network, but it adds more freedom in the definition of a relation. The underlying data abstraction would conform to the mathematical definition of an n-ary relation.
- An inter-user-relation is basically an (un)ordered set of users associated with a common object (i.e. a predicate). This predicate could be anything that can be signified with a URL (like a category, an article, a weblink, a discussion, a forum, a video... whatever). A relation might also contain links to other relations.
- Relations should be freely definable by users (like creating a group and inviting others in).
- There should also be some automatically generated relations, like common involvement in discussions, common involvement in articles... and maybe even edit wars ;-)
- Relations could also be used to create a web of trust, they might even include GPG or RSA data.
A simple weighted 1:1 relation might be "User-X edited N articles that were also edited by User-Y", or "User-X replied to N discussion topics that were started by User-Y". Another interesting example of weighted 1:1 relations could be "User-X deleted N words written by User-Y". An example of a relation that in fact constitutes a group of users (N:1) could be "Users X,Y,... are interested in physics-related topics", or "Users X,Y,... work professionally in psychology".
By allowing for a relation's predicate to be any kind of complex web object, these predicates (like "work professionally in X") could actually reach a high level of credibility by being backed up with peer-reviewed evidence and by having a "User-X trusts User-Y" relationship.
Trust is another very important factor that currently works invisibly and intransparently. This could be changed by allowing people to express their trust for someone else.
The whole concept of relations develops the biggest potential when combined with Proposal:Visualize_revision_history_in_a_fancy_way
- Inter-user relations can be used to get insights about users' behaviour.
- Most importantly, they provide the basis for improving the perception of what an article actually is: A common text edited by interconnected people. Many people may not actually grasp how significantly a wiki (like wikipedia) is actually bound to the people involved with it. By visualizing relations in a useful way, this perception may be greatly improved.
- This can also support the reader's evaluation of an article (in terms of trust, significance, correctness, affiliation...)
- How can relations impact users' perception of an article or other users?
- How do relations gain significance?
- What relations do currently exist between people and how do they fit the data model?
- What measures must be taken to prevent abuse of the concept?
- How is privacy affected by the data generated through relations?
Probably a major programming effort, especially when it comes to visualization. Developing the data model might also take some time, as it needs to interface efficiently with the database and the web frontend.
inspired by Proposal:Add_video_game-like_features
And it may help with:
Want to work on this proposal?
- .. Sign your name here!