Designing Video-Based Interactive Instructions

Prior Publications

This dissertation is based on papers published in previous ACM conference proceedings: the MixT system was published at UIST 2012 [46], DemoWiz at CHI 2014 [48], DemoCut at UIST 2013 [50], and Kinectograph at CHI 2013 [44]; DemoDraw will be published at UIST 2016 [47].
While I am primary author on all publications and led the described projects, this research could not have been completed without my advisor Bjo¨rn Hartmann and my collaborators that I have been fortunately to work with. Specifically, Dr. Mira Dontcheva and Dr. Wilmot Li at Adobe Research provided valuable guidance on three projects (MixT, DemoCut, and DemoDraw); Dr. Steven M. Drucker and Dr. Bongshin Lee at Microsoft Research guided the DemoWiz project with their expertise on visualization; Professor Daniel Vogel at University of Waterloo greatly contributed to the DemoDraw project. A group of MS and undergrad students at UC Berkeley and Adobe contributed to implementation, design, and user study in the projects, including Sally Ahn and Amanda Ren in MixT, Joyce Liu and Jason Linder in DemoCut, and Derrick Cheng and Taeil Kwak in Kinectograph.

Chapter 2 Background

This dissertation proposes computational methods to create interactive tutorials for software ap- plications and physical tasks. To ground our work in existing practices and principles, in this chapter, we define the terminology commonly used by tutorial researchers and online communities (Section 2.1). We survey research studies and literature about the motivations of people creating, sharing, and consuming tutorials (Section 2.2). Finally, We discuss the common techniques and current practices of creating instructions (Section 2.3 and 2.4). These insights are obtained from the following resources:

• 
  Studies of tutorial authorship and learning [206, 205, 212, 28, 208].
  
• 
  Guidelines for creating instructions [122, 55].
  
• 
  An online survey of 2600 individuals across DIY communities [128].
  
• 
  An analysis of 600 comments to web-based tutorials [131].
  
• 
  Books and papers of principles about visualization [209, 57, 4], technical instructions [153, 192], and user experiences [92, 36].
  
• 
  Formative studies from our research projects [46, 48, 50, 47].
  

1.