Triggers, as described in Chapter 4, are ordered lists of actions. Triggers are the mainstay of interface manipulation. It was our observation from Immunology and INDIE 1.0 that most interaction in INDIE was simply showing new groups of widgets, showing answers, playing movies, and showing new questions in response to user clicks. Our experience with INDIE 2.0 bore that out, almost.
Authors universally expressed satisfaction with triggers, as they gave them lots of flexibility in how the interface worked. One author, a student who knew others were going to take the same program as he had, said he was going to recommend INDIE to others precisely for its flexibility over other model-first tools they could use (Advise and Run).
Still, triggers were a simple laundry-list of things that could happen in INDIE in a fixed order. An important question to ask about triggers is if the abstraction of a simple list of responses worked--did authors find that these were the basic operations they needed in the right order, or did they find them too simple and restrictive?
One way to discover this is to see if authors had to frequently make use of next-triggers. Next-triggers allow authors to do two triggers at once. If authors never used next-triggers, they were able to fit their interface needs inside a single trigger, which means that triggers gave a useful set of functionality in the right order. If authors frequently had to chain triggers together in response to one event, they found that triggers were too small a unit of interface response (they needed to do more per event) or that the things they could do were in the wrong order--e.g., triggers show layers and then show a movie. If authors frequently used two triggers instead of one because they wanted to show a movie and then show a layer, triggers were too restrictive or were the wrong abstraction.
Nutrition, the largest GBS built with INDIE, had 77 triggers that had a followup trigger out of 798 overall triggers. Thus, the vast majority of the time, authors could do what they needed to with just one trigger.
Interestingly, although there are 77 triggers that needed ``next''-ness, there are only 25 actual triggers that happen after a first trigger. Thus, authors are re-using triggers. For example, students use a tabbed notebook in Nutrition to construct their report. Each tab shows a different page of the notebook, but they all have the same introduction video. The trigger on each tab ``button'' shows a specific layer and then uses a generic next-trigger to show the guide video. Every notebook tab used the same next-trigger.