Fisheye views involve magnifying a portion of a display. Typically this is used to allow a large scale of information to be displayed with easy access to detail.
Research over the last two years has shown that the expansion of a target area facilitates movement with a mouse to that target area (McGuffin and Balakrishnan, 2002). Recent work from SURL (2003b) has shown expanding targets can benefit eldery users as well.
This work is about an implementation for Mozilla based browsers in the eXtensible User-interface Language (XUL). The current implmenentation is a JavaScript builder from OPML files. Several examples are seen below:
As your mouse moves down the menu, a zoom of the menuitem text follows your mouse, decreasing as the menu item gets farther away from your current location. As seen below, this results in a shorter overall path to any menu item, while maintaining the absolute lower boundary.
In a recent usenet discussion, one commenter suggested that the target moved as you moved toward it. In fact, as shown in the above example, the original active region remains but it expands upward as you approach. Nonetheless, a subjective impression of the target moving might damage the menu's effectiveness.
Recent work suggests that mouse movements are sensitive to changes in target size while the mouse movement is occurring. This suggests a mouse movement pattern that involves successively smaller mouse movements toward the target.
Doesn't the expansion slow down getting to the final menu items?
Gutwin (2002) showed an impairment for targetting due to fisheye distortion but in a case where size and proximity of the magnified objects differed significantly than in the cascading menu list. They implemented a velocity sensitive magnification and found it performed better. I'll experiment with velocity sensing in Exscade, but assuming the visual impact is not negative, the expansion may still speed the final stages of mouse movement by making the target larger, even if it's upper bound remains the same as it was when the movement initiated.
Cascading menus
Only limited work (SURL, '03) cascading menu selection. One common error is using a diagonal path to a flyout menu and inadvertantly triggering the previous or next item in the root level while enroute to the target in the 2nd level. The expansion of the current element results in a lower probability that this error will occur.
Fisheye views are useful for solving screen realestate issues and selection from a large number of alternatives. See the HCIL Java demo.
References
Bederson, B. (2000). Fisheye menus. ACM UIST 2000
Symposium on User Interface Software and
Technology. p. 217-225. [link] [context] [Java Demo]
Byrne, M. D., Anderson, J. R., Douglass, S., & Matessa, M. (1999). Eye tracking the visual search of click-down menus. Human Factors in Computing Systems: Proceedings of CHI 99, 402-409. Reading, MA: Addison Wesley.
Gutwin, C. (2002). Improving focus targeting in interactive fisheye views. Proc. CHI 2002. p. 267-274. [PDF]
McGuffin, M. and Balakrishnan, R. (2002) Acquisition of
Expanding Targets. Proc. CHI 2002. p. 57-64.
A number of experimental interface works have begun to explore expanding target style systems, though often with gratuitous and detracting effects.
Try the Ripple Menu from Brendan Dawes.
