Technical Report Number
As computer size continues to decrease and new user interface technologies become more ubiquitous, the conventional keyboard and mouse input interfaces are becoming harder to design into newer machines and less practical for use in some applications. The pen is one input technology more suited for the upcoming generation of smaller computers using direct manipulation interfaces. However, a pen-only user interface relies on continuous gesture and handwriting tecognizers that are often slow, inaccurate, and error prone for command and text entry. Speech recognition is an input modality that can input commands quickly and potentially be a fast text entry mechanism, but lacks the capability of direct object manipulation and has inaccurate recognition. The combination of both pen and voice input should complement each other for direct graphic manipulation applications. This thesis compares the speed, usability, user-friendliness, and accuracy of a pen-only graphical editor against a pen-with-speech graphical editor.
Two versions of a graphical editor were developed which have the same functionality. One is controlled by pen input alone and the other is controlled by both pen and speech input. The pen-only editor used the tool bar for command entry and character handwriting recognition for text entry. The pen-with-speech editor used speech recognition for both command and text entry. In a pilot study using both editors, 13 computer science graduate students were asked to draw a petri net, a state diagram, a flowchart, and a dataflow diagram. Shape entry was facilitated by automatic shape recognition that transformed continuous drawing information into a perfected shape. Experimental results comparing the editor's user interfaces were then analysed. Results show that the addition of speech made the editor slightly faster. Experimental subjects claimed this editor was more usable, perceived to be faster, and preferred to use. About half of the subjects found the editor with speech not to be more user-friendly than the pen-only editor. The accuracy of character recognition for the pen-with-speech editor was significantly inferior to the pen-only editor's handwriting recognition. The low recognition accuracy was caused by the speech recognizer's inability to distinguish between similar sounding letters.
Schmidt, Karl E., "A Pilot Study of Speech and Pen User Interface For Graphical Editing" Report Number: WUCS-96-17 (1996). All Computer Science and Engineering Research.