Automatically generating personalized user interfaces with Supple
Download 5.78 Mb. View original pdf
|
| Participant { 1 . . . 17 } Each participants completed 3 × 3 × 5 = 45 trial sets fora total of 765 trial sets (270 for able-bodied and 495 for motor-impaired). The dependent measures were: • Widget manipulation time captures the time, summed overall operations in atrial set (including errors, spent by the participants manipulating individual widgets. It was measured from the moment of first interaction with a widget (first clicks or mouse wheel scroll in case of lists) to the moment the widget was set to the correct value. For many individual operations involving widgets like buttons, tabs, and lists (if the target element was visible without scrolling manipulation time resulted, because the initial click was all that was necessary to operate the widget. 942 K.Z. Gajos et al. / Artificial Intelligence 174 (2010) 910–950 Fig. 33. User interfaces automatically generated by Supple for the font formatting dialog based on three users individual motor abilities. The interface generated for AB was typical for most able-bodied participants small targets and tabs allow individual lists to be longer, often eliminating any need for scrolling. MI could perform rapid but inaccurate movements therefore all the interactors in this interface have relatively large targets (at least 30 pixels in each dimension, at the expense of having to perform more scrolling with list widgets. In contrast, MI could move mouse slowly but accurately and could use the scroll wheel quickly and accurately this interface therefore reduces the number of movements necessary by placing all the elements in a single pane, at the expense of using smaller targets and lists that require more scrolling. • Download 5.78 Mb. Share with your friends:
|