C2MS13 — Consider Using UML Primitive Types Instead of Defining Types in C2MS

From Elisa during AB Review of C2MS 1.2:

“It's odd that there's only one external reference (href) to a UML primitive type in the whole model. There are several primitives masquerading as Classes in a Package called Type (line 696) which includes classes named String, Boolean etc. This is really isn’t right, and should be corrected to reference the standard UML PrimitiveTypes (which are not Classes). Some of the classes in the Type package are disguised datatypes e.g. U32. Others are odd such as "Dependent upon response" (line 729) (does not appear in the spec) and "Variable". See sect 6.3.5 of the spec which even starts with the statement that they are datatypes (not Classes). That includes the statement that Variable is deprecated but that's not indicated at all in the XMI. Typed Variable has a structure so should be a uml:Datatype not a PrimitiveType. The primitive types should be mapped to XSD types using the XMI tags”

C2MS Response at the time:

"I agree, but we have tried to maintain backwards compatibility except in cases where the prior definition is simply incorrect (and therefore unworkable). Our approach has been to live with the way C2MS 1.0 laid out data types and structures with the intent to revisit them in a 2.0 future release when backward compatibility doesn’t tie our hands. With that in mind, my recommendation is to leave this as it is in the spec and model as long as we are in the 1.x baseline. "