With over 1.8 billion websites online in 2018, up from just 0.8 billion three years ago, it’s apparent that the growth of available information far outpaces our ability to digest it. Search engines can be of help, but only when we can form well-targeted questions such as ‘how many pints are in a gallon,’ or ‘what are the admission stats for a college.’ However, we often try to learn with little experience and might not know enough to realize the pertinent questions. In these cases, search engines alone are unable to offer much clarity on how we should or could learn.
I ran into this challenge when I first set out to learn 3D computer animation online during my freshman year at Brown University. I roughly knew what the end product could look like — perhaps a small section of an 3D-animated film like from Pixar or Dreamworks — but not knowing much more specifics about the field, it was hard to plan my learning. If my goal is to create a minute-long film about magical tree elves, would this article about keyframing or that tutorial on texturing take me 20 percent, 2 percent or 0.2 percent towards my goal? I eventually went through an 8-hour survey course on Lynda.com, which at a surface level taught me what the main areas of computer animation are — modeling, shading, lighting, animating, etc. after which I had a basic familiarity with some softwares and key components. And more importantly, I learned about what I should learn.
After more weeks of watching tutorials, trying simple projects, and wandering through help pages, I gained more idea of what project goals were feasible and how different resources could be helpful in what areas. Looking back, I realized that I could have learned more quickly and efficiently if I had known which tutorials to use, which projects to undertake, which topics make up the bulk of this field, and which are more niche details, better left for later.
So, what if there was a way to keep track and share not only the knowledge, but also the path of exploring knowledge, comprised by the connections I navigated through during my learning process.
Suppose I decide to create a map of quality content from my animation exploration. First I save all the URLs of the web pages I studied and organize them into an annotated map; I then explain each of the topics at a high level.
Now, when someone, let’s call her Jane, wants to explore how 3D animation works and maybe learn how to create a small project, Jane can glance at my map and quickly gain an idea of the main topics, some recommended tutorials, and also a few articles to get started.
As Jane navigates through her learning process, she can also sketch out her exploration in learning a topic and add more to the map. Imagine if this map were public for anyone to contribute to — as more people find resources that fill in gaps in the initial map, a collective map of resources could eventually develop.
As the map grows with more and more connections, it could become challenging to discern relevant nodes. How could we keep our collaborative map from becoming too overwhelming, so people can still navigate through it to discover quality content?
To address this, I propose a system with two scoring mechanisms. One is active, where user gives feedback to determine which nodes remain in the map over time. The other is passive and affected by user activity i.e. node usage. Nodes and paths with more user reinforcement are displayed more frequently with more prominence in the map. Exactly how much activity or user feedback (measured in upvotes) will impact the “node score” will need to be carefully designed and tested, perhaps with a “self-learning” capability to optimize its algorithm.
Node score will also decay as time passes. If nodes do not gain enough user upvotes or attract user activity, they will gradually decrease in score, be displayed less frequently, and eventually fade out completely.
The organizational structure I propose is simple — this map is constructed as a graph of nodes, where a node either contains a web URL or a list of more nodes. For instance, CG Animation would contain nodes “Modeling,” “Shading,” and “Lighting” among others, and the “Modeling” node would contain nodes with URLs to modeling tutorials and guides.
Between nodes, links can be annotated to represent different relationship:
- A Parent/Child link connects a subtopic to a parent
- Related links connect related topics
As people map more topics and areas of knowledge across the internet, we will gain a complex yet meaningful collective map. Together with smart search functionality, this system can provide an comprehensible glance on any topic that’s represented on the internet. This can be a new breed of encyclopedia, where content is highly-distributed and hosted everywhere, larger topics breaking down into smaller, interconnected subtopics so that when we consume information on the internet, we can zoom in and out to navigate rather than jumping from page to page, and to discover the best paths for our own targeted learning.
This map will be a living system that grows and learns from all users and the connections they make. It can help us navigate through and benefit from an ever-growing collection of knowledge.