1) I don’t se the benefit of individualizing Applications/OS, networks, Hardware. Why not saying “enabling generic services”. Anwyway we don’t really now how many hardware are in play, some of them could very well be virtual anyway.

2) You place SCORM in right column and xAPI on the left. While xAPI could do more for social learning it is really a replacement of SCORM, both should not appear on a diagram unless it shows transitions.

3) The pipeline/stack figure is fine to have a first idea but IMO the truth is that most modern system will incorporate xAPI feedback back into the agregation level (that’s what we do) so the whole thing from assets to management is more like a wheel rotating reentring times before output finaling get consumed. So you will end up have xAPI both above and below content.