22 May 2012

Standards Working Together: An e-Learning Perspective

Written by: Steve Lay

I have been involved with the technical standards community for many years and was first introduced to HR-XML over ten years ago. I was working on standards for learning systems (with a particular interest in assessment) and the various consortia representing different industry sectors were just beginning to colaborate to prevent duplication of effort. I later worked, briefly, as the liaison between some of the working groups of the IMS Global Learning Consoritium and HR-XML.

During those early liaison discussions it was clear that everybody involved understood the value of re-using work from other bodies to create coherent packages of standards for their own special interest group. This doesn't just save time but also adds value to those adopting the standards. For example, e-Learning developers may not be familiar with the concepts defined in a typical HR system so they need the project teams writing e-Learning standards to work through these touch points and adopt a sensible profile for their industry.

Although one often hears the standards joke, "so many to choose from", the reality is that most standards are a synthesis of existing standards combined with some unique added value appropriate to their specific audience.

My personal perspective is not one of an HR-professional but of someone who works in the related world of e-Learning. As the world of e-Learning reaches out to HR-XML so I catch glimpses of the way the HR-XML consortium itself reaches out to other bodies. For example, the 3rd generation of the HR-XML standard draws on the work of OAGIS which deals with such things as "supply chain management, general ledger operations, and manufacturing". I know little of the challenges these aspects of business face but I know that by profiling their work, the HR-XML consortium ensures that I don't have to.

At Questionmark, we work with a range of standards bodies, including HR-XML, to promote a set of robust and complementary technical standards that together underpin our Open Assessment Platform initiative.

This week, in Chicago, the next meeting of the AICC will be co-located with a meeting of the HR-XML consortium. This is an important time for the world of learning standards because the AICC have recently opened up their work on the next generation of e-Learning standards for public comment, called CMI-5.

The new CMI-5 initiative [http://www.aicc.org/joomla/dev/index.php?option=com_content&view=article&id=158&Itemid=34] aims to replace the existing, and related, SCORM and AICC standards for integrating learning content.

Both of these standards have served the community well but the success has largely been about content interoperability and the reporting of summary scores. The current impetus is to expand the range and types of evidence that can be collected and analysed.

Questionmark is also certified by HR-XML through our support for reporting results using the HR-XML data model. The advantage of supporting the HR-XML model is not the interoperability 'out of the box' that cusotmers expect with AICC or SCORM but a more subtle alignment of our data models with those of the HR system.

As I complete a week of co-located meetings it is interesting to compare and contrast these two approaches. The new work by the AICC presents an opportunity to take some of the formal data modelling from HR-XML and investigate whether or not it can be adopted directly by the e-Learning community. That might even mean literally, by inclusion of elements drawn from an XML namespace developed or profiled by HR-XML. One of the major changes being suggested in the latest CMI-5 work is an XML binding for delivering services over SOAP. Even if direct inclusion of definitions turns out too be too hard -- the AICC also have compatibilty goals -- closer alignment of the data models in areas of common functionality is a worthy goal in itself. For example, value-space definitions might be referred to directly rather than duplicated. Even a published cross-walk between two data models can be very valuable to developers and will make interoperability much more predictable (as well as speeding up implementation).

Personally I'm a big fan of this pix-and-mix approach to developing specifications. To me, the nouns defined by the HR-XML specification provide an entire library of data elements that can be referenced in the standards of related industries - their usefulness goes beyond the suggested scenarios and pre-defined messages. In fact, it is worth getting an idea of scale here. The work targetted by CMI-5 is probably about the size of just one of the HR-XML scenarios incorporating or just one or two nouns.

Reusing existing resources like this allows new specifications to be developed more quickly, and to focus on the added value they bring. That should also result in more timely publication and improve adoption.