Computable content: Lessons learned

Description

Paco Nathan shares lessons learned about using notebooks in media and explores computable content that combines Jupyter notebooks, video timelines, Docker containers, and HTML/JS for “last mile” presentation, covering system architectures, how to coach authors to be effective with the medium, whether live coding can augment formative assessment, and the typical barriers encountered in practice.

Paco outlines the difficulties presenting interactive learning environments—in everything from tutorials at conferences to live online courses to relatively novel forms of on-demand content, in subject areas spanning software development, systems engineering, data analytics, business soft skills, and design thinking—he encountered as a media and event producer.

One example: how do the authors of a popular open source framework teach 500 people in a conference tutorial, each with their own live coding environment? Installation and configuration alone may take up to an hour—half of the tutorial’s scheduled time.

Paco then walks you through the key questions O’Reilly Media considered for addressing these challenges through computable content, including:

• How can we make learning materials more powerful by integrating compute engines and data services?
• How can we coach authors and instructors to be effective with the medium?
• How can media production techniques use video as subtext?
• Can live coding augment formative assessment?
• What theories support this pedagogical approach?
• What are typical barriers encountered in practice?
• What kind of system architectures are needed at scale?

Paco then outlines some solutions and discusses the impact of notebooks for sharing and learning across an organization. In one solution, each user session launches a Docker container running on a Mesos/Marathon cluster through a gateway built with nginx and Redis. The UX is entirely browser based, for a fully personalized compute environment that can cover a wide range of programming languages. By leveraging Docker, additional frameworks and data services can be added; by leveraging HTML and JavaScript on the frontend, there is zero installation required. The solution is also instrumented for data collection and analytics and is being extended to offer live coding within a formative assessment platform.

Parts of this software have been released as open source on GitHub.