First pass feedback #6
Description
Just finished reading the xCaliper document. There is an amazing amount of detail here and you clearly have thought deeply about the specifics of the transduction. Here are a few thoughts for consideration:
- I think you are intentionally taking a stronger competitive tone with this document, but be aware that some of the language (e.g. Assertion Added introductory functional and market context #1) may come off as too critical. Whereas some of the prior conversations have been along the lines of bringing the two specs together, this is clearly a recommendation for Caliper to subsume xAPI.
- Some may object to characterizing IMS' approach as "Pay to play". Maybe "member-supported"?
- I would not characterize Kaltura as a content provider per se; move of delivery platform.
- This is a key paragraph: "If xAPI terms, profiles, and recipes could be mapped and then deprecated in favor of associated Caliper profiles then both the IMS and the xAPI communities can join together to work collectively on expanding metric profiles to cover more and more learning activity use case scenarios." This formulation aligns the two communities to a common goal and I would move closer to the start of the doc. If you can establish agreement that both communities should be working to build relevant metric profiles, not competing over data structure and transport protocols, then you may be able to argue for a quicker semantic "convergence" of data streams (but not specifications).
- Here you do a nice job of showing the two specs coming together vs. simply competing, which may be more palatable politically: "We should also consider extending the Caliper Event model to include certain xAPI concepts such as Statement Authority, Statement reference (i.e., a pointer to a pre-existing statement), SubStatement (i.e, prediction or statement of intention, or anticipated future action), Statement attachments and perhaps Statement signing. Caliperizing these concepts will help maintain fidelity during the conversion process and reduce the need to unmatched xAPI concepts to Event.extensions."
- the terms staterecipements seems like a typo?
- "duck typing" may need some explanation.
Overall, great work. Keeping all this in your head for a single spec is impressive but doing with two and cross-walking them is an amazing achievement.