squaremobius dot net

Reasoning in Zowie

From the magical red notebook, scribblings:

  RDF model :—
    Choose what relations mean: linked? 
                                contained?
                                ignored?

So, you have all these semantic relations between chunks of things; you decide what display relations correspond to the semantic relations. The display relations are things like navigation hierarchy, spatial containment, linking.

  In the ZUI, the display of relations determines what
  gestures mean.

I.e., if you drop an item into a box you are asserting a semantic relation between the item and the box. This is nothing but direct manipulation. These can be contextual as well; for example, if you create a link between a text fragment and [something representing] a person, you might be asserting a 'creator' relationship; but, make a link between a text fragment and another text fragment and you are building a narration.

This might fix a fundamental problem with the idea of multiple views (I want to keep Zowie's model honest, so the idea is to plan to have multiple views from the outset). The problem is this: You can't influence things important to one view from other views. For example, spatial layout in the ZUI is (likely) missing from a Web site view. When the dimensions are choosable, though, it doesn't matter so much. The spatial layout matters only so far as it implies semantic relations — to-the-right-of might mean narrative succession, for example.

I am making a hash of explaining this. I should just write some code.