Suggested High Level architecture

It recognises two distinct domains: 1. The mobile app 2. What you need back home at HQ The mobile app: - Would need to be a native app in order to package content - Seeing as a single tablet will likely go to lots of kids, a key requirement is for the kids to identify themselves (and there ought to be incentives for doing so) otherwise, you can only track usage and progression data at device level. In other words, you¹d need to think about the data you want to generate and design the solution in line with that. - In order for you to be able to generate xAPI data and capture this on the device (which will likely be offline most of the time), any learning content would need to be packaged within the app. I.e. I don¹t think it¹s possible for one app to generate data for another app on the device without an internet connection. - I therefore propose that any content would need to be developed as HTML packages (pretty much the regular scenario for e-learning packages) and delivered via an embedded browser (which is straight forward to achieve). - this likely means that you will want a content creation framework so that you have a degree of consistency and can develop content cost-effectively. A consistent interface is also likely to make the learning easier as the kids will get used to the standard interface. This is where a project like Adapt can help: https://community.adaptlearning.org - the app would have to have a content catalogue - I.e. A selection mechanism for which content to launch. Given that the kids may not be computer literate, an icon-based, graphical interface may well be the best approach. The icons would need to be considered in the cultural context - e.g. How best to express a subject like Œmaths¹ etc. - the app would essentially work in the same way as some elearning Œoffline players¹, I.e. They carry the content on the local device, can launch it within a browser and Œfake¹ an API, which the elearning talks to and sends tracking statements to (this is what is meant by ŒxAPI connector¹ in the diagram. There would also need to be a database / data store for the statements attached to the app. - while the devices are expected to be offline, I still think it¹s worth building in the ability to configure, sync and download data remotely if there is an internet connection. Presumably, it is easier to get the device to a place where there is connectivity take connectivity to the devices (e.g. Via satellite phone) in some circumstances (even if not all the time) - this is a Œjust in case¹ thought. For this reason, I have also included the content sync and data uploader components in the concept. Systems at HQ: - there would need to be an LRS into which all the data can be poured. A key question here is the level of granularity to which the data attaches - I.e. Is it Œone record set for the whole project¹, one record set for each device or one record set for each identifiable user. - there is an open source LRS called Learning Locker, which may help: http://learninglocker.net/, https://github.com/LearningLocker - this has a built in web frontend for admins, which does some basic analysis or at least data visualisation - to this master data record, you could attach analytics software and create reports. Some of these you may wish to make publicly available - there would also need to be a data uploader service (whether this is a web service, which you can call remotely or a script, which allows you to import data copied directly from the device) - over and above this, I think you could think of how to enable partners to create content for the project - so a content creation tool and a shared content catalogue might be worth considering. I hope this makes some sense - I know it has been a while since we talked. Please feel free to ask any questions you may have. All the best, Sven