![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
usernamenumberBecause the form had character limits, I didn't get to submit exactly this, but I'm pretty happy with what I wrote, and since it's a project I'm genuinely excited about, I thought I would share.
Federated learning experiences: toward the invisible LMS
The modern e-learning world includes a wealth of innovative and powerful tools for developers of educational content. Some of these are built into learning management systems (LMSes) like Sakai, Moodle, and edX, others are designed to interact with LMSes using standards like SCORM and xAPI, and still others are designed to operate as stand-alone learning experiences, with little or no thought given to integration with a particular platform.
This talk starts with a central premise: “the more an LMS tries to do (e.g. creating content, hosting content, recording grades, etc), the more limiting it ultimately is, resulting in lock-in, inflexibility, and lost options for innovation and experimentation.” From there, it proposes taking a page from the anti-monolithic philosophy of Unix: each component of a system should do one thing very well, and cooperate well with other components for more complex tasks. In other words, authoring and hosting content, storing student records and providing a user interface that ties it all together should be handled by independent, modular components of an organization’s e-learning ecosystem.
This is illustrated with a demonstration of Swag, a non-profit project being developed by tunapanda.org. Swag “federates” learning experiences from an arbitrary set of sources, which can include stand-alone content like XOT objects, content hosted in one or more traditional LMSes, and things it would be difficult or impossible to integrate into a traditional LMS, like activities using non-web-based applications. These are presented using a game-like interface that organizes experiences into “maps”, and awards points (“swag”) for each accomplishment. A set of scripts gathers learning records from each source, converts them into a common format (xAPI), and stores them in an independent learning record store. The result is an approach to learning management that in its most radical form could enable completely self-directed, self-motivated learning with minimal dependence on any particular platform for creating or hosting educational content.
This is all open-source, and if you have the time and inclination to help (it's mostly node.js and Django), get in touch, and/or check out the git repo here:
https://github.com/tunapanda/swagmaps
I also submitted this:
Using Ansible to automate the maintenance of your e-learning ecosystem.
Ansible provides an easy but powerful way to define configurations that should be applied to servers, and can make deployment, maintenance, and change control easier whether you have a small number of servers or hundreds. This workshop will demonstrate building a set of Ansible “playbooks” to provision a Linux virtual machine hosting Xerte Online Toolkits and its dependencies (web server, database, etc).
This one also touches on a really cool open-source/non-profit project I've been working on, which aims to use ansible to make it easy to provision classroom servers that can provide a local wifi network with locally hosted educational resources from sources like wikipedia. edX, and Khan Academy with no dependence on an external Internet connection, for deployment in facilities where bandwidth is either non-existent or prohibitively expensive. We even have a test deployment in a refugee camp in Jordan coming up soon.
http://github.com/tunapanda/provision
Federated learning experiences: toward the invisible LMS
The modern e-learning world includes a wealth of innovative and powerful tools for developers of educational content. Some of these are built into learning management systems (LMSes) like Sakai, Moodle, and edX, others are designed to interact with LMSes using standards like SCORM and xAPI, and still others are designed to operate as stand-alone learning experiences, with little or no thought given to integration with a particular platform.
This talk starts with a central premise: “the more an LMS tries to do (e.g. creating content, hosting content, recording grades, etc), the more limiting it ultimately is, resulting in lock-in, inflexibility, and lost options for innovation and experimentation.” From there, it proposes taking a page from the anti-monolithic philosophy of Unix: each component of a system should do one thing very well, and cooperate well with other components for more complex tasks. In other words, authoring and hosting content, storing student records and providing a user interface that ties it all together should be handled by independent, modular components of an organization’s e-learning ecosystem.
This is illustrated with a demonstration of Swag, a non-profit project being developed by tunapanda.org. Swag “federates” learning experiences from an arbitrary set of sources, which can include stand-alone content like XOT objects, content hosted in one or more traditional LMSes, and things it would be difficult or impossible to integrate into a traditional LMS, like activities using non-web-based applications. These are presented using a game-like interface that organizes experiences into “maps”, and awards points (“swag”) for each accomplishment. A set of scripts gathers learning records from each source, converts them into a common format (xAPI), and stores them in an independent learning record store. The result is an approach to learning management that in its most radical form could enable completely self-directed, self-motivated learning with minimal dependence on any particular platform for creating or hosting educational content.
This is all open-source, and if you have the time and inclination to help (it's mostly node.js and Django), get in touch, and/or check out the git repo here:
https://github.com/tunapanda/swagmaps
I also submitted this:
Using Ansible to automate the maintenance of your e-learning ecosystem.
Ansible provides an easy but powerful way to define configurations that should be applied to servers, and can make deployment, maintenance, and change control easier whether you have a small number of servers or hundreds. This workshop will demonstrate building a set of Ansible “playbooks” to provision a Linux virtual machine hosting Xerte Online Toolkits and its dependencies (web server, database, etc).
This one also touches on a really cool open-source/non-profit project I've been working on, which aims to use ansible to make it easy to provision classroom servers that can provide a local wifi network with locally hosted educational resources from sources like wikipedia. edX, and Khan Academy with no dependence on an external Internet connection, for deployment in facilities where bandwidth is either non-existent or prohibitively expensive. We even have a test deployment in a refugee camp in Jordan coming up soon.
http://github.com/tunapanda/provision
no subject
Date: 2015-02-10 05:33 pm (UTC)Also, if you want stories of people trying the monolothic approach (badly) I heard of one team that was trying to build their LMS within salesforce....
no subject
Date: 2015-02-10 05:49 pm (UTC)It was a mistake inasmuch as that courseware means I am locked into Moodle even if a better option comes along, because my content and user records are all stored non-portably in a monolithic LMS.
It was *not* a mistake inasmuch as they're good tools that I've used to create some good content, I just don't want that to mean I can never try out edX, Sakai, or any other potential alternative, nor easily use tools like H5P that don't offer good integration with Moodle.
This is where the idea of the federating "meta-LMS" comes in. Imagine an interface like the skill tree in an RPG, where each "skill" is a learning object that can be hosted in anything that generates or can be made to generate xAPI statement. I can keep my Moodle stuff in Moodle as long as I write a script that queries Moodle's database to extract the records I care about as xAPI statements that go to the LRS that Swag reads (and that's only in the absence of actual xAPI integration, which they're working on). Other nodes on the map can be hosted in edX, Sakai, or anything else that generates or can be made to generate xAPI statements.
LTI is also a step in this direction (though I've seen way too many implementations that treat it as just a way to embed content from one LMS into another, with no facility for actually transporting scores or other data), but from this is more radical idea with where the record store, content host(s), and connective interface (the "skill tree" in our implementation) are all completely independent and modular.
no subject
Date: 2015-02-10 06:52 pm (UTC)But yeah, taking it to the next level and dis-aggregating the user profile and the data that attaches to it would enable a more portable and/or modular system.
RPG skill trees are a bad source example, because they exist within the context of a single RPG. Fine as a UI metaphor though. Some of the work done around skill badges (Mozilla Backpack) is probably a little closer to this idea. Or digital health records, where the individual owns the data on more than one level, but there are standards and ways of exchanging some or all of this information amongst providers in such a way that they can add their pieces and access the data they need to provide a service...
To a certain degree this is the old Department Store vs. Boutique argument. But if we don't have reliable transportation, sometimes the choices are made for us.