What is a "Platform for e-Learning"?

There has been a lot of talk in LETSI lately about what SCORM 2.0 should look like. Everything is on the table. Since there are so many use cases for e-learning, it has been difficult to narrow down the focus of what a specification for the next generation of e-learning should look like.

Some have proposed that maybe SCORM 2.0 should be more of a platform than a specification. Aaron Silvers introduced the phrase “Linux for learning” (Aaron discusses it some here). It has been bounced around a lot, there is a lot of enthusiasm around the idea, but it’s exact definition remains a bit elusive. There is a thought that in a time of accelerated innovation, it makes more sense to focus on informal specifications and open source implementations rather than on formal standards. If that is the case, what is LETSI’s role in the community? Does LETSI merely shepherd the development of specifications, or do they actually produce open source software?

It is sometimes hard for me to debate ideas in the abstract. It really helps me to see the merits (or demerits) of an idea if we can be concrete about how it will actually be implemented and what it will do. In this case, what does a “platform for e-learning” actually do? What problem(s) will it solve? What will it look like? Why would people actually use it?

Tim and I have batted these last questions around a lot over the last few weeks. We’ve come up with one answer that might represent a useful path forward for LETSI. This might be just a different way of stating what others have already stated, but here goes.

Ten years ago, every LMS and every content vendor needed custom hooks to get their applications to work together. Proprietary interfaces required vendors to modify their products every time they wanted to integrate with another vendor. The world looked like this:

After SCORM standardized the interface between LMS’s and content, things got a lot easier. Now LMS and content vendors all just had to code to one single interface and things were a lot simpler. The world looked like this:

Today we are back in a similar situation, except the items being integrated are changing. The content integration problem has largely been solved (yes, it can still improve, but the path is known). The integration problem facing learning systems now stems from trying to integrate other systems and applications. It looks something like this:

It’s a very similar problem to the one SCORM solved. Years ago, LMS vendors were being asked to integrate with many different types of content. Today, they are being asked to integrate with many different types of applications. Currently, LMS’s integrate with things like virtual worlds, simulations, assessment engines, competency management tools, content repositories, reporting services, discussion boards, classroom management tools, intelligent tutoring systems, performance support systems, knowledge management systems, document management systems….and the list goes on and on.

Back in the day, each LMS and each content producer had a proprietary interface for performing integrations. Today’s situation is similar. Today, many LMS’s have their own proprietary plug-in architecture or API for enabling extensibility. For a great example of LMS extensibility and integrations, check out the number of modules available for Moodle.

Many LMS’s are providing extensibility interfaces. Each is doing this differently. The industry wants more integration. The use cases presented to LETSI require system integration and extensibility. Perhaps standardizing LMS extensibility interfaces is the next big problem that LETSI should solve.

Integration and extension seem ripe for standardization. Why?

  • It is currently being done. As a standards body, we would not be innovating, but instead codifying established best practices.
  • It is in demand. LMS’s are already expected to integrate with a number of systems and this number is rapidly increasing.
  • It is currently costly. Much of the cost of LMS implementations currently stems from integration services work.
  • It is currently painful. In addition to cost, the fragility of integrations makes it hard to justify changing systems. This creates vendor lock-in.

Many have already discussed developing a “platform for e-learning” or a “learning systems architecture”. This post is saying the same thing, but getting a bit more concrete about what that means. To us, this platform/architecture is a way of adding new functionality to existing learning systems in a standardized fashion. A standardized platform/architecture enables functionality to be developed once and integrated into any system, just like content can now be developed once deployed to any system.

The key insight here is that this is something that is already being done and that it is an existing pain point with real business value. LETSI can define an architecture by codifying existing practices rather than inventing something from the ground up. Simpler integration benefits vendors and users alike.


  • philip

    Well-stated, guys… I second that motion.

    I think standardizing LMS extensibility — essentially standardizing a kind of LMS plug-in architecture — would solve many problems in our industry and would go a long way towards enabling developers to implement new technologies as they become available.

    This also isn’t far removed from Aaron Silvers’ suggestion that the portability of *data* is what’s important here.

    One concern would be how do you standardize extensibility when the LMSs use a variety of programming languages (ASP, PHP, Java, etc.)?

  • Allyn J Radford

    As a participant/co-chair of the architecture committee that is contributing to the next “SCORM” or whatever it may be eventually be called, I really appreciate Mike’s post. It raises some very pertinent issues and gives a different opportunity for these types of issues to be discussed.

    I would like to make several comments in relation to the specific issues raised because they are meaty issues. First, in LETSI’s meetings in October in Pensacola there was a significant focus on what was loosely termed, the “out-of-browser” experience. As Mike indicates, there are a lot of applications that are currently deployed for different aspects of technology mediated learning, and only some of those are browser-based and a sub-set of those are LMSs. The learning world is less and less about what can be achieved by integrating with an LMS and the learning experiences available therein, to a much broader set of experiences in serious games, simulations, clients other than browsers etc etc etc.

    Second, one could also argue that initiatives such as the Open Knowledge Initiative (OKI – http://www.okiproject.org) based at MIT is one potential solution to contribute to the sort of single plugin or API approach to integration spoken about here. It is not perfect but it provides significant utility and flexibility and it already exists. If that is the solution being sought, then I would argue it is better to identify those initiatives that already exist and divert attention to making them easier rather than starting again. We already have wheels… Let’s aim for something else, even if OKI is not the exact answer.

    There are several elements that I think are consistent in all the messages I have seen and heard.

    a) service-oriented architectures are as necessary for learning infrastructure as any other infrastructure (or piece thereof).

    b) content is an important part of learning but we want to use it differently. soa (service-oriented approaches – to use the e-framework differentiation at http://www.e-framework.org) change the game for content. It now needs to be consumed by different applications and communities want adaptability at the authoring and aggregation stages not just interoperability at the publishing stage. These are quite different to support.

    c) Learning life is more complex, richer, more diverse, more distributed, more community-based (as the differentiations), and are provided through an increasing set of applications (wikis, blogs, LMSs, simulations, games, off-line materials, online spaces etc etc).

    There are a range of other issues but this comment is rapidly getting too long – it is a big issue.

    Mike is 100% correct. The key question that LETSI has to answer before it can proceed is that given the range of initiatives that are currently contributing to this space, “What is the piece that LETSI should do with ‘SCORM’ 2.0 and what does it mean for the learning, education and training communities LETSI already serves and/or is seeking to serve?”

    Apologies for the length of the comment but I hope it contributes to the framing of the debate and the solution.

  • Damon Regan

    Great thoughts as usual Mike. I appreciate your desire to get concrete. I posted on my blog a stab at the question raised here: “What is the piece that LETSI should do with ‘SCORM’ 2.0?

  • Dan Young

    Mike – I think you hit this exactly on the head, at least in reference to the “out-of-browser” use-cases that we have been looking at. The SCORM RTE provides a channel for learning content to communicate with the LMS for status data, learner data, control data, etc. We don’t have any equivalent API for apps, unless the app can somehow pretend it’s a SCO.

    I think one of the pieces we need a new model to provide is the ability for the learning event to initialize/launch the “LMS” rather than vice-versa. Currently, the LMS has to launch SCOs, whereas a real out-of-browser learning/assessment event might not be launched or initiated by the LMS at all. SOA would allow the event to initiate the transation with the LMS (or whatever we’ll call the record-keeping software function). Just having that basic capability would open up a lot of possibilities.

    One suggestion, though. Rather than think of apps as plug-ins or extensibility for the LMS (which strikes me as a pre-SOA model implying a kind of parent-child relationship), I think we should be thinking instead of loosely-coupled peer applications, where either one could be considered the server or the client, depending on the nature of the transaction. Just a shift in terminology and/or perspective.

  • Mike Rustici

    Phillip: I agree with you on the similarities with Aaron’s comments on the “portability of data”. That’s what this interoperability is all about. I considered trying to represent data and its various formats in the diagrams I created, but my artistic imagination failed me! The problems of what data to exchange, in what formats and and what quantities might seem a bit staggering at the moment, but I imagine it felt that way when people first started to look at the problem of standardizing content interoperability as well. When you start to look at what is unique to learning however, I think the problem becomes a bit clearer. People in the learning industry like to think that we are quite unique, but I disagree. When you really start to look at it, the technical aspects of learning technology aren’t hugely different from the rest of the software world.

    Yes, platforms and programming languages are certainly diverse across systems, but I think this is where a loosely coupled, services-oriented architecture comes in. In a standardized world, “plug-ins” wouldn’t be as tightly integrated (from a technical perspective) as they are now.

  • Mike Rustici

    Allyn: Thanks for your feedback. There seems to be a lot of great discussion happening related to LETSI on independent blogs and various email threads. I wonder if we should find a better way to structure these on the LETSI site. The current wiki seems a bit clunky for the purpose though.

    I agree that the learning world is less and less about the experiences delivered from an LMS, but I do still see a need for those experiences to be tracked and managed by an LMS. Also, note that I use the term “LMS” almost figuratively. While we are focused on the learning world, the “systems that track and manage learning” are rapidly evolving. They are even morphing/merging into other types of systems. LETSI’s work shouldn’t be tied to today’s concept of an LMS. BUT, I think it is very important to start from someplace we know and to start from someplace where there is a current market need.

    If OKI has laid a suitable foundation, then by all means, let’s take a look.

    I agree with the messages you point out this thread as being consistent with. Before starting the post, I was thinking to myself, “am I really saying anything new here?”. I don’t think there’s new ground-breaking material, it’s just a different slant that crystallizes the abstract notions and frames the problem in a way that is easier to pitch on an elevator and makes good business sense.

  • Mike Rustici

    Dan: I agree the master-slave relationship is probably the wrong model going forward. I think how it’s presented though might vary based on whether your audience is business focused or technically focused. As a geek, I’m with you, “loosely-coupled peer applications”. As a business man, that sounds mystical and it sounds like something new, something that is going to make me change, something that is going to cause me pain. I think it’s important from a business/marketing perspective to be saying “we are going to standardize something that you are already doing to make your life easier and save you money”. From a technical perspective, we can know that we are going to make it better and cooler.

  • Mike Rustici

    Damon: Your thoughts are similar to some of the things Tim and I talk about internally. Going back to my response to Allyn’s comment, we consider the big difference between “learning content” and just “regular content” to be three things: “context”, “path” and “assessment”.

    “Context” – Some form of orientation as to how this information fits into a broader picture.

    “Path” – A structured and intentional way of presenting content so that knowledge is built upon and opportunities for new discovery are presented.

    “Assessment” – Ensuring that knowledge is being properly conveyed to enable the path to be adapted.

    We’re certainly not experts in education by any means. Those are just our thoughts as technologists working in this area.

    Damon’s post is at http://damonregan.blogspot.com/2008/12/thoughts-on-letsi.html

  • Nathan Ashlock

    I’ve been contemplating e-learning application interoperability for several years now. Back in 2001 I was involved in the development of a whitepaper that discussed an alternative to SCORM which focused on systems in a sense sharing students and not content and proposing the development of a common ‘language’ used to communicate a student’s interactions with training content while the student was ‘visiting’ a remote application. It was also based on the reality that one SCORM conformant system cannot effectively train everything just because it is SCORM conformant (gaming and simulation applications for example).

    We did not get the opportunity to take the concept beyond the whitepaper level and start developing the common ‘language’ models. However, the K-12 education industry has somewhat tackled this problem with a standard called the Student Interoperability Framework (SIF), available here: http://www.sifinfo.org. I think the SIF specification can be extended to fit needs of platform for e-leaning model.

  • Pingback: LMS Open Source LMS - 10 alternativas a Moodle | e-Learning para el nuevo milenio()

  • E-Learning

    Nice article, however LMS is now becoming an somewhat old concept. There are new alternatives out there, which are becoming incredibly useful to more and more companies.

  • Pingback: LMS Open Source – 10 alternativas a Moodle « Wiki – Tu Academia.com()

  • Pingback: SCORM » The SCORM Engine as a "platform for e-learning"?()

  • Pingback: SCORM - "Which part of SCORM should I implement in my LMS?" » SCORM - ()

  • Pingback: niche jet()