Re: FW: XML Schema datatypes: NaN, lists of union types,[NEL],misce llaneous editorial

[Eddie Robertsson <eddie@allette.com.au>]

"Arnold, Curt" wrote:

> Thanks for the clarification, but I do think on this point Section 2.5 is still unnecessarily misleading and inconsistent with the cos-list-of-atomic section that you referenced.  Sorry, I haven't
> lived with the document as long as you have so my comments were based strictly on the prose.  I glad that the intent was not to bar lists of unions.
>
> "list datatypes are always derived. The value space of a list datatype is a set of finite-length sequences of atomic values. "
>
> Okay so far, since each individual item in the list is ultimately an atomic value.
>
> "The lexical space of a list datatype is a set of literals whose internal structure is a white space separated sequence of literals of the atomic datatype of the items in the list (where whitespace
> matches S in [XML 1.0 (Second Edition)]). "
>
> Unfortunately, here it uses "the atomic datatype" which implies an exclusive atomic datatype shared by all items in the list.  If the phrase was "the atomic datatypes" or "the atomic or union
> datatype" then things would be okay.
>
> "Definition: The atomic datatype that participates in the definition of a list datatype is known as the itemType of that list datatype. "
>
> Again "the atomic datatype" is used.

I asked these same questions some weeks ago and Henry pointed me to the correct place [1]. One other thing that Henry pointed out is that a list type cannot be derived from a union type that in turn is
derived from a list type. I haven't found this information in the specification and if this information isn't there you could _indirectly_ derive a list type from another list type (via a union type).

Cheers,
/Eddie


[1] http://www.w3.org/TR/xmlschema-2/#derivation-by-list