Our user testing on TaxonTree did indicate that undergrads at least enjoy browsing when there is animation and optional zooming. However, I'm concerned at the speed of the animation in this particular implementation -- there is a risk of disorientation even though the layout is more predictable and retrievable than a hyperbolic tree. Personally, I found it a little difficult at first to distinguish the children with different parents at the same level. DOI does help. We used DOI (among other techniques) in CandidTree, an app. that is being presented tomorrow at Interact 2007. Our focus was on comparing trees, and I guess I can't say we tested with "normal folk" as our subjects were either curators or Microsoft employees :-)
Cyndy
Rebecca Shapley wrote:
Yes, it's a degree of interest interaction. There have been various attempts to implement this DOI idea since it came out, and I'm pointing to this one as a rather novel implementation, that may overcome some of the usability issues seen with hyperbolic trees, etc.
I would definitely test its usability with a taxonomic tree data set and some "normal folk" before advocating it as the ultimate solution for an EOL project. Denise Green and I did these types of usability tests on three interactive tree presentations available in 2005, providing something of a baseline against which this particular interaction style might be judged. See: http://groups.sims.berkeley.edu/TOL/docs/GreenShapleyTOLFinalReport.pdf (4 MB)
-R.