Learning from comments on YouTube

December 18, 2011

Web video is playing an increasingly prevalent role in online learning: university open courseware initiatives, flipped classrooms experiments, and a growing number of virtual schools all rely heavily on video for direct instruction. And while video is itself a broadcast medium, situating video within a website opens the possibility for peer-to-peer interactions. Most video-hosting websites support some peer interactions through a shared comment stream, but this inherently treats the video as a two-dimensional object, ignoring its most meaningful dimension: time. Particularly for long-form presentations of complex concepts that are difficult to grasp simply by viewing, a venue for peer dialogue is incredibly valuable, but a venue that lacks a notion of temporal context is inherently limited. What could peer interactions around web video look like if temporal context was treated as critically-relevant? I’ve been lucky to have the opportunity to explore this question over the past few months while building Grockit Answers.

While YouTube, Vimeo, and several other video-sharing services provide programmatic access to a hosted video’s timeline, the new technology that really opened up the possibility of interacting with the temporal dimension of web video is Popcorn.js, a JavaScript event system for HTML5 media. Grockit Answers structures peer interactions as a set of Question & Answer discussion threads, each one anchored to a particular point in a video’s timeline. When a viewer asks a question, perhaps about something specific that they found confusing, Popcorn is used to automatically time-stamp the question to associate it with the particular moment in the video at which it was asked. When others viewers reach that point in the video, the question appears on the screen, presenting it at the exact moment that these viewers have the same context as the person who asked the question. Discussion is no longer about a video, it can now be scoped to a particular comment, explanation, or demonstration within a video. My absolute favorite part of this design is the often-surprising result of watching a video that has been the subject of many questions: As you start to become confused by a complex or vague explanation, the questions forming in your head pops onto the screen. Perhaps even more surprising, answers to your questions – perhaps clarifications or alternative explanations – also start appearing. While this is an asynchronous peer interaction, the time-triggered displays give the feel of just-in-time live help. I’m calling this a pseudo-synchronous interaction.

Many thanks to the Mozilla Popcorn project and the Popcorn.js development team for providing the layer of abstraction that makes this type of learning environment such a joy to build. For another great example of new forms of social learning around web videos, check out Popcorn Macbeth and some of the other great Popcorn demos.

By Ari Bader-Natal

Software Carpentry » What I’ve Learned So Far wrote on December 19, 2011
[...] Allow in situ commenting. [...]

Why I'm excited about Khan Academy Computer Science :: twosigma wrote on August 20, 2012
[...] Learning from comments on YouTube [...]