As a Mentee I want to receive suggestions for the next most relevant topic(s) for me to learn when I complete a topic

[bsenyk]

No description provided.

[jmatsushita]

The content could be added to a queue (as in "next in the queue" or a "playlist") implemented as some kind of tagged linked list, or a linked list of maps.

For instance someone having just completed the safe-social-networks topic and the chat unit (socnets:facebook-chat) would have the following queue

index:	0	1	2	3	4
group:	basics		intermediate		advanced
practice:	socnets:getting-started	socnets:who-controls	trust:social-networks	threats:social-networks	opsec:identities
howto:	socnets:facebook-chat	socnets:facebook-*	socnets:compromised-accounts		socnets:alternatives
tech:	internet:basics	internet:traces	threats:social-networks-tech		socnets:installing-alternatives

• After completing a unit, the user should be presented with:
  • What the current "next" recommended topic is based on our adaptive engine (at index 0)
  • The possibility to change this to another topic (see below)
• A general principle should probably be to give the learner agency over whether:
  • To go "back" and learn what is necessary to understand or to provide background and knowledge about what the user's intention is.
  • To "stay" and review more equivalent content.
  • To go "forward" and learn more advanced topics.
  • Other axis could be
    • "deeper", "shallower"
    • or "more in the details", "more in general terms"
    • or "more in practice", "more about how it works"
• One example for this could be:
  • I consume content that is of the category protect:content for a while. The adaptive engine does its thing.
  • I start consuming content in the protect:identity category, the adaptive engine adds this as a new "layer" and should start by making sure I have all the basis covered in the protect:identity category before moving on to more "advanced" topics.
• Some form of learning requirement dependency might need to be implemented. See https://kumu.io/iilab/open-mentoring#open-mentoring for ongoing attempt at this.

[bsenyk]

It seems like an MVP of this would be to simply include links to related units in the content itself as metadata. These could be enhanced by on-device data such as what has already been read or what's relevant based on settings. Until we have a large enough repository of content that is classified in the above manner, we should hold off on anything more complicated.

[bsenyk]

See #15 (a lower-priority ticket) for adaptive learning

[jmatsushita]

I've started working on the content dependency (see updated Kumu graph) which will be usable as metadata to add new content to the "queue".