July 21, 2016
8 PDT/9 MDT/10 CDT/11 EDT/16 BST/17 CEST
Attending: Ellen Meiselman and David Topps, Co-Chairs; Rob Chadwick, Valerie Smothers and Luke Woodham
1 Review minutes
The minutes were accepted as submitted
Ellen suggested modeling the MedBiquitous registry off the Tin Can API registry, which allows people to browse profiles. This could exist on medbiq.org in the near term. David liked Jason’s ADL registry that crafts profiles and links to RDF data structures. His concern was the timing of when that might be available and in the meantime, they are getting pressure to move ahead. Ellen commented that we would have medbiq.org IRI’s no matter where the profiles are hosted. Github would be appropriate for version control of profiles or definitions of verbs. Valerie asked for clarification of specifics on version control. Ellen provided the example of the specification for Tin Can on Github, where someone is in charge of making changes to controls, version numbers and releases. Luke agreed with Github. The versioning is not an issue there but how to manage persistence. Valerie recommended adding this to the medbiq Github site. Rob added Github was successful for managing changes to document but not to manage URL’s. David suggested following up with Jason on the timeline for the registry database and his degree of confidence in that timeline. Valerie agreed to follow up with Jason.
3 Update on Open Labyrinth development
David provided a summary of the workshop he and Ellen held at the annual meeting. They stress tested a volunteer using an Open Labyrinth case while connected to Arduino machine tracking stress levels while learners rapidly made clinical decisions with minimal information. They brought together data from the camera, Arduino, Open Labyrinth, and HP5. They now have a tool to manipulate parameters to generate xAPI statements. David was grateful to Andrew Downes who did interesting work with the data. Ellen commented you could visualize stress levels with color.
David questioned how to pull data sets together and know that the session and learner are the same. Within context you can specify registration UUID, but one device or source must generate it and propogate that to the other systems. Rob commented that he had surveyed LTI, CMI5, and Tin Can Launch to address that problem. He proposed a simple algorithm, available on GitHub, taking the best parts of each. One system allocates the token and delivers to the content. The token can be exchanged with the launch server. The XAPI endpoint is responsible for attaching context to activities. Cryptography was used to ensure content was correct.
Ellen asked what the flow would look like for a student. Rob replied that one system needs to act as the portal. The student logs in and clicks a link which contains token and URL. The content uses that to get a new token (could be endpoint). Cookies are established to maintain the session. The system is allowed to post to the end point until the terminate command is sent. The student can’t come back later. The approach is similar to Shibboleth. David agreed that was a sensible approach. Brian at Tin Can suggested using CMI5, and its ready now. Tin can launch also kind of works. Rob mentioned CMI5 has a token handshake. With Tin Can API, students can impersonate the system. ADL launch provides the opportunity to encrypt the token using the key of the content. ADFL’s approach is not dependent on an LMS. If ding client side HTML, the user must have the private key.
Rob volunteered to set up a demo that could be part of a future meeting if the group was interested. Ellen suggested a private working meeting between Rob and David and then further exploring with the group. Valerie will email Rob for his availability, possibly for the September meeting. Ellen suggested David post notes and ideas to the wiki. Valerie will investigate the BAA announcement regarding funding from ADL for simulation and team-based activities. Valerie asked if applicants can be from outside the US. Rob didn’t know the logistic specifications but offered Valerie to follow-up with Dr. Damon Regan. Valerie will contact him following the call.
- Valerie will follow-up with Jason on the timeline for ADL registry availability.
- Valerie will email Rob regarding his availability for a demo.