This needs to be figured out by trial and error, but here are some preliminary criteria:
- We should be able to define (and see) context of any statement in the structural way (not by plain text). So this has implications to the structure -> we need to have some kind of contextual relation between entities
- It should have a way to approach disagreements (handle alternative views, and allow to support each view, and calculate the “rightness” of each view, coming to a conclusion about the prevailing point of view) as one of the driving factors of content creations
- We should be able to navigate data in different dimensions: such as, we should be able to “wrap” and “unwrap” back the statement through logic with UI tools. We should be able to tweak the context to see how the statement changes.
- I thnk that v1.0 of the UI should allow for turning on/off user ID. On the stage of design we should be able to see who sais what for better meta-communication (discussing the actual design of the system), but we should also see how things work with IDs off.
- I don’t think “ease of use” should be part of the v1.0 interface :) It just doesn’t seem plausible. A raw technical model should be built and tested first.
- A big question is how to handle branches, interlinkage. Need to see google.wave’s experience.