> I think the idea has merit. > > Am I correct in implying that the "sensitivity" would differ from say > the classification attributed to a particular reference document?
Yes. The idea is that "sensitivity" - or perhaps more appropriately, "classification" - would apply to the model itself, not the resources that the model was derived from. > I effectively include a "sensitivity" measure in the AERO-ML data files > I create using a "variableDef", but a more formal approach may provide > better consistency when exchanging datasets. Similarly for the > "licence". > > In addition to 'name/type', would it be worth having 'refID' and > 'description' attributes associated with licence and sensitivity (or > what ever they may become)? > > Geoff Brian I think that at least for the "license" element it might be appropriate. One can list the license type as being "unlicensed", or "GPL", or some other unique license. But it is probably a good thing to list where that license is published, either online or an address where the license could be obtained. I don't know whether that location should be element content or an attribute. Jon