Paper Reading 10 - Gesture Search: A Tool for Fast Mobile Data Access

Comments
Patrick Frith
Paola Garza

Reference Information
Gesture Search: A Tool for Fast Mobile Data Access
 - Yang Li
 - UIST 2010, New York, New York

Meant to be an "A," but
could also be an "H."

Summary
This paper was about a new search tool called Gesture Search.  As its name implies, the user draws gestures instead of typing to search for items.  This particular version was designed for Android phones.  The user simply draws a letter with their finger on the screen.  Using handwriting recognition software, the program deduces which letter possibilities the gesture could be, and displays a list of results that could relate to its deduction.

The initial results and the
next letter.

There is a timeout feature that allows for separation of drawn characters, since only one can be drawn at a time.  The more gestures a user gives, the more specific the results become.  In order to become a more efficient tool for the user, the program learns the user’s search habits.  If they continually pick the same result after the same gesture, this result will float to the top of the results list.

The final results with "An."

The designers performed a usability test with a large group to find out if the program was in fact quicker to use than standard typing searches.  They found out many useful things during the study.  One was that the most popular items searched for using Gesture Search was contacts.  They also found that 74% of the gesture queries did not require any deletions.  This showed that Gesture Search was very accurate in finding what the users wanted without the need for typing.  And lastly, the vast majority of gesture queries only required two gestures or less.

Search history influencing
future searches.

Discussion
This seems slightly similar to the stroke-based typing discussed in my last paper assignment.  I like this gesture direction of tools.  It is similar to how programming continues to drift farther and farther from machine code.  The more abstract the code writing process becomes, the easier it is for humans to fully express their ideas in code form.  I think the reason these gesture-based methods of searching and typing are so efficient is because of how abstract they are.  With typing, we adapt our ideas to the tool.  With Gesture Search, the tool is adapting to how we behave and is drastically simplifying and speeding up the search process.  This is the second paper in a row that makes me sad I don’t have a touch screen phone.