Wednesday, March 13, 2013

#LSCON Experience API and LRSs

Some really nice conversation with Glenn Bull from Skilitix before Aaron's session.

Seems like there are a lot of us grappling with learning analytics and its place in the enterprise.
Not just us end-user folks.
------------
Experience API more like an open standards project.
Define and make instruction set available.

Baseline info on Experience API (Tin Can)

Creating structured record jSon object
Actor- verb -object -result -context

Learning record store - place where activity statement goes.
Records any valid statements
It is a capability - not a product. Make across systems.
Try to use same command across all systems that can read common API.
Want to be able to leverage LRS idea across varying types of systems. Project management, linked in, enterprise systems

Verification done through signed statements

LRS addresses only 1 capability that LMS deals with. The learning records.
LRS = antenna
LMS = fortress

LRS collects. Reporting system makes sense of the data.
The collected data can go across to other systems - Hris systems, LMS reporting, even a personal data locker (maybe through LinkedIn?)

LRS = storage not processing.
Very little validation done at runtime.
Other services make meaning of the data.

Identify person talking about through authentication means.
Can use embox (?) email address. O auth - token that tracks back. Like Facebook or twitter signin from another app.

Can do both group and team (will need to figure out how to architect . This is still early days.)

There is a spec for activity streams.
Google etc care who you are.
Learning folk care what you are actually doing.
Verb list describes what you are actually doing. (Fire a kiln vs. fire a gun)
URI -creating a unique identifier for that verb.
Objects described in lots of ways. LRMI - trying to describe learning resources in common ways.

Result - did you get a result back?
This done separately. (Also traditional thing we measure)
All activity time steamed.
Working on geo-location.

What are people actually using (can we leverage this with existing web analytics?)
How can we build in feedback loops - how is what you are doing (training) translating to the job?

(Good stuff Aaron!)

No comments: