Ask Slashdot: Best Practices For Using a Reputation Engine To Rate Information?

GrantRobertson writes: For my graduate project, I am considering developing a web engine designed around sharing and organizing actual information in a way that people would actually like to and easily be able to use it. Unlike a wiki, the information will be much more granular with lots more metadata and organization. Unlike a web forum, the information will be be organized rather than dispersed throughout thousands of random posts, with little room for dominant personalities to take over. While I like Stack Overflow, I am planning far more structure. While I enjoy the entertaining tangents on Slashdot, I don't want those to take over sites created using my engine. Naturally, there must be some way to prevent armies of bots or just legions of jerks from derailing web sites created using this engine. Given that, what would you say are some good rules to include in the reputation engine for such a site. What kinds of algorithms have you found to be most beneficial to the propagation and spread of actual knowledge. What would you like to see and what have you found to be dismal failures?

Comments

sounds like OneModel; maybe you can start there

[lcall]

Sounds like what I'm trying to do here (AGPL): OneModel.

It doesn't have all the features, but what you describe is partly there, or planned for the future, though for now it's in the form of a text-only UI and you have to install postgres. The UI is something like a mix of git's "commit --interactive" and gopher (remember that, anyone?), but it is very efficient if you just read the screen and are a touch typist. Probably currently most suitable for someone who now uses emacs org-mode, or collapsible outlines of any kind, but wants to handle richer kinds of information (eg, GTD...) and a more task-specific UI.

It's what I use as my own personal organizer and knowledge manager, but ~"sharing" features for collaboration, including reputation and others, are on the wish/plan list. Feel free to use it as a starting point, or join the list for discussion. I was hoping to get the web site updated with a later binary and an enhancement, and much more information on my future plans, by roughly next week. It still lacks a convenient installer but the INSTALLING file in github is current.

If interested you could always get on the announcements list for when I add features. My health isn't great at the moment but I hope to be able to sell binaries or installers in the future for part-time income or the like. Patches or discussion on the list are welcome. I have been thinking hard about this since about 2000 and am glad to finally have something others can use, though the potential audience will be larger once there are better installers and other needed features, UIs etc.