I recently returned from IMS Global Learning Consortium’s 2014 “Learning Impact” conference held in New Orleans. IMS Global is a non-profit member organization that develops interoperability standards like “Question and Test Interoperability” (QTI) and “Learning Tools Interoperability” (LTI) that are widely adopted in the education sector.
While standards (yawn) don’t get most of us out of bed in the morning, there’s definitely a buzz around IMS Global’s activities at the moment. Institutions from kindergarten to higher education are increasingly requiring compliance with various interoperability standards. This was evident in several of the conference panel discussions. What’s more, startups see interoperability as opening markets dominated by the big LMS vendors. Publishers want their content to be “plug and play.” Venture capital is pouring into the education space and this tends to generate buzz.
At Flatirons we’re best described as “standards agnostic.” We’ve long experience with standards from a variety of domains and industries. We’ve done DocBook, DITA, and S1000D. We’ve seen standards success stories as well as major fails. Sometimes interoperability is really needed and a standard is the only way. Sometimes a standard represents free intellectual property – someone has thought through the hard information modeling semantics for you. But blind adherence to poor standards, or when they serve no purpose, can be a big waste of time and money.
IMS Global’s standards run the gamut. QTI is one of the oldest and most challenging of those standards. We’re currently working a large project for high stakes assessments using a severely stripped-down version of QTI 2.1. For that particular project, we wanted to leverage the information model of QTI where we could, but interoperability wasn’t a primary concern. It was interesting to hear similar stories from several other conference participants – production uses of QTI generally result in a stripped down version of QTI.
IMS Global would likely respond that QTI is intended not just to model question types, but rather all of the complicated structures for answer rubrics, scoring etc…. It’s not intended to be a “render format.” But in trying to do everything in a gigantic schema (and additional Schematron rules), they’ve lost most of us. Perhaps it should be more than one standard, or we need standards, or at least well-known paths, to go from QTI to a render format, and vice versa?[1]
Several conference sessions looked at EDUPUB. EDUPUB is best described as an umbrella standard for packaging and delivering educational content. As IMS Global puts it, EDUPUB is
“…a comprehensive model for the interchange and deployment of educational content based on the W3C Open Web Platform and expressed as an integrated set of specifications including EPUB 3 (IDPF), LTI, QTI (IMS), LRMI and other emerging standards.”
The EDUPUB sessions included an interesting look at the work of Thor Anderson, Utah Valley University department chair and IMS Global contributor. He demonstrated a student project that used various EDUPUB standards to plug Readium,[2] an open-source Java ePub 3 reader, into Moodle and play QTI quizzes packaged as JSON. The demo used a JSON “render format” for QTI along with AngularJS on the front end to manipulate the QTI-based JSON.[3] While derived from QTI, the JSON manipulated in the UI was clearly a radically simplified subset of the QTI constructs. It’s not clear from my recollection of the sessions, but I’m hopeful that EDUPUB may start to define a standard way to produce browser-friendly JSON[4] from QTI.
If QTI is the old-school bugbear of assessments, LTI is IMS Global’s new, cool kid. LTI defines a relatively simple interoperability standard.[5] For example, an LMS can link out to a third party learning tool. LTI manages the passing of student information, authentication and the launching of the new tool. LTI has gained wide adoption in a relatively short time and has enabled new integrations and business models amongst tool makers, publishers and LMS vendors. Widespread dissatisfaction with the idea and practice of a monolithic LMS has contributed to its popularity.
A highlight of the conference was Vital Source architect John Tibbetts’ sessions on LTI 2, the next generation of LTI. LTI 2 further streamlines the linking of applications and tools. According to Tibbetts, LTI 2 has gone beyond just interoperability to being an “LMS-neutral application platform.” This is a fascinating idea and one we’ll explore here at Flatirons as we develop architectures in education.
The conference concluded with an informative panel[6] of “money guys” – venture capitalists and private equity leaders in the education sector. Several panelists (though not all) agreed that the way to succeed in the education sector was with “platforms.” They seemed to think that monolithic platforms are needed in order to provide a great user experience and to do comprehensive analytics. This runs counter to the conference theme around standards (e.g. LTI) enabling more heterogeneous learning environments. Another standard, IMS Global’s new Caliper standard is pioneering a model for analytics in learning such that analytics from a variety of tools and platforms can be aggregated and will share the same semantics. Several panelists seemed to think that effective analytics was only possible by using monolithic platforms where all analytics were coming from the same source.
The monolithic platform approach also runs counter to our experience in a variety of markets where proprietary, one-size-fits-all solutions almost never seem to create customer satisfaction and value. Large platforms are expensive and, due to a lack of focus, are most often not up to the challenge of creating a quality user experience in all aspects of the system. Our bias these days is towards standards-based platforms and a mix of best-of-breed, third-party components. We always advise our clients to get into the market quickly, see what works and what doesn’t work, refactor and repeat. An agile, lean approach to getting solutions into the market presents less risk to the clients and more value, more quickly to their end users.
The view that investment is best made in comprehensive platforms, that looks to bet on the next dominate player, is similar to the Silicon-Valley ethos of big ideas and “disruptive” new technologies.[7] Disruption and homogenous platforms run counter to evolutionary, collaborative nature of standards bodies and the open-source communities that have sprung up around them. While as a consultant I’m bound to be practical in adoption of standards, I’m personally rooting for the collectives – the standards and open-source communities that are crafting elegant, focused learning software that will truly engage learners.[8]
[1] Going from a render format like JSON back to QTI is required for browser-based authoring tools that produce QTI.
[2] Readium came up in another session by SoftChalk. Sue Evans of SoftChalk said that Readium was the best option they had found for delivering ePub 3 to browsers and to Android platforms.
[3] In addition to using the Caliper standard to instrument the assessment engine for analytics, the demo included a simple, AngularJS-based QTI authoring tool. Technically the authoring tool was creating JSON constructs that could then be transformed to QTI. The project saved the JSON locally in PouchDB and eventually synced to CouchDB. IMS Global is planning to release this POC code to members in late May or June.
[4] For example, IMS Global has started to use a more structured form of JSON “JSON-LD” in LTI 2. http://www.imsglobal.org/lti/ltiv2p0/mediatype/application/vnd/ims/lti/v2/toolproxy/id+json/index.html#The_JSON-LD_Context
[5] It’s worth noting that LTI is standardizing a much different, simpler domain than QTI. LTI is about plumbing, integrating tools. There is a much higher level of complexity in modeling all aspects (e.g. rubrics, scoring) of assessments.
[6] Panelists included Roger Novak, General Partner, Novak Biddle Venture Partners, Brian Napack, Senior Advisor, Providence Equity, Michael Locke, Former Vice Chairman, Rasmussen, Inc., Manoj Kutty, CEO, LoudCloud Systems Adam Newman, Managing Partner, Education Growth Advisors, Matt Haldeman, Vice President of Technology Partnerships, McGraw-Hill Education
[7] An early general session featuring Ben Nelson, founder and CEO of the Minerva Project, was a fascinating look at completely reinventing the very idea of a university.
[8] I noticed several IMS Global contributing members working on code for standards demos during the conference sessions.