Statements instead of a vocabulary

The newer approach records what happened. Making sense of it is your problem.

The more recent specification in this field abandons the fixed set of tracking fields and instead records statements of the form: somebody did something to something, with whatever additional context the sender chooses to attach. It can describe a simulation, a book read offline, a task performed on a machine, an assessment observed in a workplace. The expressive range is genuinely much wider.

The trade is that the receiving system is no longer obliged to understand the statement. Two organisations can both be conformant and produce data that cannot be compared, because they chose different vocabularies for the same activity. The specification anticipates this and provides for shared profiles, and profiles work exactly to the extent that people agree to use them.

This makes the choice less about which is newer and more about what question is being asked. If the requirement is that a course reports completion to a system of record reliably, the older approach is a better fit and remains completely serviceable. If the requirement is to capture learning that happens outside a delivery system entirely, the older approach cannot express it at all.

Running both is common and not incoherent. A great many organisations use the established packaging for the courseware library, where interoperability is the whole point, and the statement-based approach for the workplace and simulation evidence that the older model was never designed to hold.

It is also worth being sceptical of the framing that presents one specification as having replaced another, which is largely a marketing construction. Long-lived standards in this field have consistently coexisted, because they were adopted for different reasons by organisations with different obligations, and a system of record does not change its data model because a newer approach is available. Planning on the assumption that the established one is going away has, so far, been a reliable way to be wrong about a decade.