I've seen some discussion going by regarding potential extensions to the exhibit syntax, but another direction in which it would be quite useful to push is in extending the set of available views. There have been quite a few interesting visualizations bound to a specific blob of data; they'll be more useful if you can feed them from exhibits. What those views tend to do is use a few properties of a data set to guide their construction; with an exhibit you could plug arbitrary properties. A few examples come to mind:
* a "tag cloud" view. a tag cloud is driven by two properties; one defines the order of the tags and the other the size of each tag. in delicious alphabetical order and tag popularity are used but one could equally well birthday to sort and net worth to size a set of people. * a graph-layout view where the items are connected by a line if they are related by a given property. More generally one could allow multiple properties and use a different color or shape of line for each. Or perhaps some animation and focus as in this tool: http://der-mo.net/05_links/index.html * a pie chart view using any number-valued property to size the wedges, or a histogram view to show number of items with each value of a given property, or some other favorite type of chart. * a scatter plot that places items in a grid according to the number values of two distrinct properties. _______________________________________________ General mailing list [email protected] http://simile.mit.edu/mailman/listinfo/general
