SCORM

Hi Martin,
as per my understanding, it should be technical enough to address some overall thoughts about Packaging and LMS vs LRS role.

I guess that most of the "pure" LMSs claiming now the Tin Can support are "just" providing a "Tin Can adapter" to:

• intercept the JSON payload
• read it
• map it into their pre-existing tracking storage, when applicable¹

Packaging is out of scope in XAPI while it is required in CMI-5 (an Application Profile adopting Tin Can) and probably by a significant part of the "traditional" Market: that's IMHO the reason why is still pending with some proposals and some Vendors already guessed among them.

That being said, it would be nice to listen for opinions - even beyond the "cold" technical spec implementation details - about:

1. In the past Test Suites have been proposed to provide self-assessment tools for Vendors and Customers: is there something already planned for CMI-5 too?
2. Could CMI-5 be a driver for no more than a partial implementation of the XAPI specs to allow LMSs provide support "just" for Learning Packages tracked via JSON too, kind of slowing down the path to a full LRS adoption and the discovery of potentially new learning paradigms (shortly: less AUs/LOs, more Learning Experiences/Pills)?
3. Learning Packages have been always part of the Market, to feed Customers' Catalogues and Specifications&Application Profiles have been the required drivers for Interoperability: while it is clear that XAPI has given freedom for more creative (AND track-able!) learning approaches, people is still looking for traditional Learning Packages too. Has SCORM 2004 really failed - probably Vendors first - to provide a well defined packaging structure where the Author has the power to fully control reusability and delivery strategies with no code but XML (complex) rules and opened the doors to just the same LOs but JSON-powered, missing² the pedagogical fine tuning already available in a "simplified" manner with the AICC and SCORM 1.2 prerequites and then (too much?) amplified by SCORM 2004?

TIA,
Matteo

¹That is, at the moment, where MDL-35433 is, though being a good starting point for thougths.
²For those interested, find here the CMI-5 course structure current proposal: https://github.com/AICC/CMI-5_Spec_Current/blob/master/cmi5_coursestructure.md#72-course-structure-xsd

Here are my notes from the CMI-5 presentation by William McDonald. I hope I can get the slides too...

"ML" indicates my personal opinion.

• Better than SCORM, overcome limitations
• all media, any media
• potentially huge courses / media
• user can upload any media, large media
• recording of events/tracking can be very granular, and largely left for the content to define granularity and meaning ~ ML: this is great but makes reporting both an opportunity and a headache
• "technical framework that different communities of practice can put more structure on" ~ML: so it punts on some aspects.
• CMI-5 is an XAPI "profile" -- maps XAPI to the traditional LMS model **launch interface **session mgmt **course structure
• CMI-5 = XAPI + "some extra rules" ** from AICC PoV, Tin Can is not quite complete, because "profiles" aren't established; so interop isn't there. CMI-5 will be the first XAPI profile, and defines a few more things so that interop can be achieved in practice. Docs at github.com/AICC - git://github.com/AICC/CMI-5_Spec_Current.git
• Aiming for end of year ~ ML: he didn't say which year

What's already defined:

• Web browser launch interface -- more planned: native iOS/Android/Windows (this covers when the LMS is a native app, and wants to launch CMI-5 content and how that content gets a handle on where the LRS is, auth, etc)
• other things I missed -- need to get the slides!
• State API -- ML:stuff the LO ("AU"!) stores in the LMS, and can be used to resume a session
• Statement API ~ ML: stuff stored for reporting
• Verbs
• Activity Types
• Move-On criteria
• more things I missed - get slides!

Scope of CMI5:

• launch by LMS
• runtime communication (XAPI)
• LMS course definitition (as needed for this)
• reporting requirements for LMSs ~ ML: actually they are not wanting to define this, prefer to leave it open
• CMI-5 runtime interface is dependent of course structure stored in the LMS (?)
• Launch API - The LMS builds a URL and calls the "content" url with some extra params that should be ignored by the server -- they are meant for the content, client-side (JS code should read the URL from the DOM and parse it).
• State API -
• Statement API - used by the AU to record progress to LRS/LMS - verbs & activity types. CMI-5 adds rules to require that statements are tied to a sessionid. ** ML pushed for a discussion -- launch API doesn't contain session ID. Martin here thinks it's a problem -- current launch dance is brittle and racey. We may be able to influence this. If we get sessionid added to the launch API quite a few things get simplified.
• Verbs ~ ML: this makes sense to me. I would join handling of abandoned and exited; mobile usage blurs the line.
• Activity types : fluid string names / descriptors ~ ML: they seem to be "static" -- this was unclear. If it is static, it should be in metadata so the LMS doesn't have to "run" the LO to know if it's a video.
• ML: Looks like the AU can store anything (pics, videos, etc), and the AU can retrieve anything it has stored. It splits "state" from "statements" so it is clear what is reportable and what's state.
• ML suggestion for reporting - add some metadata to the AU that provides hints about the datatypes that the AU will store. Perhaps JSON schema. This allows the LMS/LRS to store the data in matching DB columns for better scalability & perf.

It was a very informative session, and William indicated several times we have an open invite to join the standards committee, this is mainly through confcalls. This might be a good idea. If we can't, we can interact via email of via their "tracker", but it is not so smooth.

If we participate, we have a voice. To have a formal vote, we would need to be a "member organization". However, the voice of a developer does carry weight

My notes are a mess. Apologies...