"Learning experiences become meaningless if they cannot be attributed to identifiable individuals and/or groups"

[garemoko]

This is a bit strong; anonymous data can be useful. Just needs re-wording a little.

See http://tincanapi.com/2013/06/05/deep-dive-actor-agent/#comment-1836520796

[rklancer]

When researching whether we should develop our own custom logging and analytics, or support xAPI and learning record stores, this phrase caught my attention as a possible sticking point.

At Concord Consortium we're experimenting with xAPI for use by educational
researchers who might analyze user actions in tools like http://mw.concord.org/nextgen/ and http://concord.org/projects/codap.

In both cases, we need to be able to glean insights from the actions (clickstreams, etc) of both logged-in users and anonymous users; not only is asking users to log in a barrier to casual use, but because much of our work is federally funded research, that pertains to K12 education no less, our ability to collect data from identifiable users is more restricted than it is for private businesses.

[garemoko]

Absolutely. I don't think this sentence is meant in any way to prevent use of anonymised agents; it's more intended to explain why identifiers can be useful and just a little badly worded.

[rklancer]

Good to see; thanks!

[canweriotnow]

Just as a note, going into 2.0 the whole IFI issue needs to be revisited with extreme prejudice as we start trying to build xAPI for IoT.

[garemoko]

I hit a use case recently (accounting professional bodies) where it would have been handy to have used an activity object as the actor of the statement. The work around is to use an account object with essentially a meaningless name, and focusing on the homePage, but that is a work around.

Certainly worth raising a separate issue for this @canweriotnow that we can tag.

[hallshouse]

We could pass this on to the CASS project. Yesterday in our initial CASS Project Community meeting, they confirmed that CASS information are meant to be part of statements and stored in the LRS as such.

In @garemoko 's example above, a CASS statement can be used for the "accounting professional bodies" group. That statement (according to the CASS project) can have sub-statements assigned to it.

In the sub-statements, associations to activities and groups can be defined. With this scenario, the user can remain anonymous as long as they are assigned a group.