lecture by Andy Cockburn - speed dependent zoom. Good idea but could be improved. needed way to turn paged in digital lab book *Inspiration!* Application didn't require zooming so SDAZ interesting but not exactly want was wanted in this case. SDAZ also required certain applications. First test of SDAZ didn't work Moving to tablet so using those affordences - SDAZ desktop based. Radial scroll - when pass radial line scroll down. Nearer centre - scrolls faster, further out scroll slower. Settings - number of lines scrolled per radial lines, scrolling page or line, number of radial lines etc The future of radial scroll? dependent on user drawing circle around centre point. Works great on tablet (pen) and desktop (mouse) more difficult on smart board because of hand-eye coordination issues (drifting). Need to make it eyes free. Experiment task Needed to find headings in document and specific pictures in group Finding headings done as same in Andy Cockburn's experiment so easy to compare back to. Doesn't test everything (ie comprehension) but does test speed Main difference with radial scroll experiments was pictures and text separate 8 people in experiment Useful hints (no order) * You need training exercise on how experiment is going to run not just on how to use the software * Automate as much as possible * If possible get to the state where you don't need to be involved at all If you have reached the level you don't need to be there then you have got to the right level of the software (you still need to be there to observe obviously) * Log as much as possible * If possible electronically analyse the data especially with more people * Don't use start/stop buttons - people concentrating to hard on task they forget * Use prompt cards to show user what they are looking for if it doesn't interfere - very useful if language issues * Perform test yourself first help spot problems that might not have been obvious ie physical restrictions * Recommended to do pilot study before actual study use friend especially one who knows about testing don't be afraid to through away and totally re-jig test * Don't let task last too long. Test subjects don't like to be doing it for too long * Double and triple test check for duplicate entries or alternate understandings of requirements * Testing will take at least twice as long as you expect * Never underestimate the subjects ability to learn from previous sessions * Know how to do statistical analysis before you do the experiment * Need to double check design/protocol * If possible new document/collection with each test since people will remember stuff/gain experience * Post-experiment interview good Factors: Text/Image (no sig diff) Short/Long (better for short) Normal/Radial scroll Tablet/Smart Board Findings: * Radial scroll faster on big screen but no difference on tablet Suspected that it was because other scroll techniques are just *so* bad on smart board * Comments - people preferred radial scroll on big board as more comfortable Main problem was that people had to keeps eyes on radial area or start wandering. Need to make it eyes-free. Arc detection rather than vector detection. Will require more and more complicated maths Java not seeming to get enough mouse events People liked being able to just go around in a small area One left handed person - really liked it because scroll bars on totally wrong side for them Expert at scrollbars but beginners at radial scroll - a bit unbalanced Be good to test on PDAs since very little screen so different again