From ADL Team Member… Nik Hruska: Which Verbs Would You Use To Describe Learning?

In my previous blog post Using Activity Streams in Next Generation SCORM, you read about Activity Streams and understood how they will be leveraged in the architecture for Next Generation SCORM.  The same technology is the core behind the major social media outlets including Facebook, Google+ and Twitter, as well as emerging applications such as Hojoki, Pinterest, and Schemer.  To review, an Activity Stream is a sentence (or statement) with an actor, verb, and an object. Let’s look at a simple example:

Nikolaus Completed CPR Training

In this example, the actor is ‘Nikolaus’, the verb is ‘Completed’ and the object is ‘CPR Training’. Activities are interactions with real or virtual media and objects such as reading an article, watching a video, flying a flight simulator or playing a serious game.

Because you care that different systems report similar activities in similar ways (what we call interoperability), content developers and Learning Record Store (LRS) vendors must ‘speak’ the same language when reporting learning experiences. The use of context and ontologies will ensure that the language being used is interpreted to have the intended meaning. Here is a simplified example which could cause an issue if my eReader was reporting how I experienced Hamlet to two different LRSs:

1. Nikolaus Read Hamlet
2. Nikolaus Completed Hamlet

The LRS could be ‘smart’ enough to recognize these synonyms, but the content developer may have a valid reason for choosing either verb.  Alternatively, applications may take common verbs as inputs and output more meaningful verbs which were not part of the original framework.  This mapping responsibility of the application could reduce the load on the content developer.

The ADL Technical Team is currently researching Activity Stream verbs and objects while attempting to align them with existing learning activities such as courses, games and virtual worlds. We are looking at defining an initial set of verbs concerned mainly with consuming real or virtual media. Since we are essentially ‘inventing the future’, we made it a priority to align with emerging Semantic Web standards (RDF, RDFa, OWL, schema.org, Microformats) provided by the major search engines (Google, Bing and Yahoo). Aligning with semantic web standards will enable future applications to intelligently tailor content to individual users by enabling the search and retrieval of contextually relevant learning material – without human intervention.

Determining which specific verbs and activities describe learning should be a community-powered effort. There will be a pattern to our language – a model – that should be extensible (“future proof”). New verbs and activities can be added over time as new technologies emerge, because what follows the emerging technologies will be new ways to describe how we use and learn with them. The model must also balance simplicity and complexity. The majority of applications may only require a list of 5-10 verbs to choose from, while others may require verbs we have not yet identified (or verbs which have not even been coined yet, such as ‘Tweeted’, ‘Googled’ or ‘Pinned’).

When looking at the implementation of the (Facebook) Open Graph API, the verbs and activities are determined by the application developer. This suggests that future applications may use verbs and activities which the community did not initially identify, but which the application requires to describe the activity appropriately. For example:

Nikolaus Ran 10.2 miles

If we haven’t defined the verb ‘Ran’ in our communities’ best practices guide, an application may not behave as expected when reporting to other systems – but that may be acceptable in some cases. Some applications built for a highly specific type of training or learning activity should be able to define and use the language necessary to define their activities. For example:

Nikolaus Fired Hellfire missiles

Most learning scenarios may not require the verb ‘Fired’, but a specialized simulation may use ‘Fired’ and several other verbs that are necessary to adequately describe the interactions. In addition, ‘Fired’ can have different meanings in different contexts (ex. Nikolaus Fired Administrative Assistant). The primary consideration is the level of detail required for your assessment. Do you need to know, for example, that the learner simply scored 42 points and won a football game, or do you need to know pass completion percentage, average yards per carry and average yards after catch?

New capabilities arise when changing the tense of verbs.  Using present tense allows for formative assessment instead of just summative assessment. For example, in a football game many types of interactions occur (Center Hiked Ball, Quarterback Handed Off Ball, Running Back Ran Ball, Linebacker Tackled Running Back). These Activity Streams all report what has already happened. The game could report at a much finer grained level using present tense verbs. The time difference reported between ‘Running Back is Running Ball’ and ‘Linebacker Tackled Running Back’ gives us information on how long the Running Back possessed the ball. More detailed information could be used to PREDICT what the learner may do to tailor the training (i.e. on third down with twelve yards to go, the learner will pass 92% of the time).

Changing verb tense could also enable a mobile GPS-powered orienteering exercise to track more than just the fact that the learner made it to the specified way points. Instead of just reporting that ‘Nikolaus Found Location Alpha’, the system could report that ‘Nikolaus Is Searching for Location Alpha’ and then ‘Nikolaus Found Location Alpha’. An assessment tool could identify those learners that had large time gaps between the ‘Is Searching’ and ‘Found’ activities to determine those learners requiring remedial work on orienteering. The Facebook Open Graph API Activity Definition page shows a simplified example of reporting activities in progress (present tense) and after completion (past tense).

Another consideration we face is the level of detail required to future-proof the initial roll out of verbs and activities. If we start by lumping everything into a few high level, abstract verbs (ex. Experienced, Read, etc.), splitting into more specific verbs later may cause interoperability issues. So where do we draw the line? We’re looking to you, the users and the community, to provide insight and feedback. Tell us what you think by tweeting your ideas to us at the @ADL_Initiative or use the comment thread below.