`type` instead of `ref`? about ids

[aothms]

xs:restriction is not a built-in data type. Those are listed here in the spec https://www.w3.org/TR/xmlschema-2/#built-in-datatypes

I also want to stress the obligatory https://www.quora.com/Is-W3Schools-a-reliable-source here I don't think we should base our decisions on that source.

Also there are apparently other issues with xsd.exe handling ref incorrectly (although not very similar to this case) https://stackoverflow.com/questions/44028046/how-to-use-xsd-exe-with-attributegroup-ref I personally have mixed experience with automated codegen from XSD.

In my understanding, but I'm not an XML/XSD expert, and I know @berlotti has consulted people infinitely more knowledgeable than me on this regard, it's as you hinted on in the other issue, in this case mostly a matter of taste. When you use type, you declare something of that type but still need to give it a name local to your schema [in XML files that adhere to the schema you see the local name]. When you use ref you basically nest that thing verbatim in your schema [in XML files that adhere to the schema you see xs:restriction].

Another difference, but not relevant, but what's described in your table, is that type can only be used for, well, types. And ref is much more general, can be used for identifiable elements, including types.

[berlotti]

It is an explicit choice to use the xs:restriction as is. We don't want to create an IDS restriction of the same type as an xs:restriction, but use the actual xs:restriction. This brings a lot of added value with the ability to use multiple toolkits for xs:restriction saving implementation time and adding consistency.

Combining two namespaces was a deliberate choice. The use of two namespaces in the XSD is not common, but valid.

Closing this issue.