This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Proposition: Including the simple type definitions for the primitive builtin datatypes in the schema document for schema documents is neither necessary nor desireable. As it stands many schema validators reject the schema document at http://www.w3.org/2001/XMLSchema.xsd because of these definitions, and they aren't needed to fulfil the primary purpose of the SDForSDs, namely to constrain the structure of schema documents. In 2e we attempted to patch around the ways in which the SDForSDs violated our own constraints on SDs, but the result is arguably imperfect, and certainly doesn't contribute to the simplicity of the spec. We could just remove them altogether, or put them in a separate section of the REC and/or a separate schema document on the Web. The simple type definitions for the built-in derived datatypes are also not necessary -- should we (re)move them also? One note of caution -- we have published URIs of the form http://www.w3.org/2001/XMLSchema#string and http://www.w3.org/2001/XMLSchema#int.maxInclusive and the SDForSD contains anchors with those names. Some people might suppose that those two things are connected, but it's not clear to me that in fact they are. . .
The WG did discuss this topic earlier this year, maybe in connection with editorial proposal EP-09 (the 'literate programming' proposal). As far as I can remember, that discussion resulted in no clear direction, so I believe the correct thing to do now is to treat the question as a new item and ask the WG to classify it.
Connected with this issue is the mismatch (in the current status quo document) between the schema document for schemas and the tableau at the end of Datatypes section 4.1.6. The schema for schemas has (pseudo-)declarations for the primitive datatypes, many of which contain annotations. But the tableau in 4.1.6 says that the primitives have {annotations} with a value of the empty sequence. Either this discrepancy is a problem and must be fixed, or it is not a problem and the WG needs to decide that it is not a problem.
Decided at the f2f in Toronto 2005-11.
The Toronto draft minutes (http://www.w3.org/XML/Group/2005/11/xml-schema-ftf-minutes#id0x04998078) say: * structures.xsd: presented as an appendix, as now, but defines no built-ins * datatypes.xsd: presented as an appendix, and as an external file, as now, but defines no built-ins * XMLSchema.xsd: presented as an external file, as now, but defines no built-ins * primitives.xsd[1]: presented in an appendix and as a separate doucment in XSD syntax and described as not a schema document, e.g. this document is in XSD syntax but violates some constraints; [it's thus not usable with conforming processors, but may be useful to other processors so it's provided in this syntax for their convenience.] * datatypes.xsd[2]: presented in an appendix and as a separate doucment in XSD syntax and described as not a schema document, e.g. this document is in XSD syntax but violates some constraints; [it's thus not usable with conforming processors, but may be useful to other processors so it's provided in this syntax for their convenience]. We will review this document after the story on composition is clear. [1] I think this should be primitives.nxsd [2] I think this should be derived.nxsd Noting that these are working names, likely to be replaced before publication
Candidate change in dg sfs-1933, sent to WG for review 2005-12-16: http://www.w3.org/XML/Group/2004/06/xmlschema-2/datatypes.b1933.20051216.html
The Working Group discussed this issue on its telcons of 6 and 13 January; at the latter it sent the issue back to the editors with instructions to revise the proposal so as to remove the xsd:annotation elements from the XML representations of the built-in datatypes, to accord with the WG's preference to have the {annotations} property of those types be absent (or, alternatively, the empty sequence). Accordingly, I'm changing the status of this item from needsReview back to needsDrafting.
On 20 January 2006 the Working Group considered, amended, and adopted a proposal to move the XML source declarations for the built-in datatypes out of the schema for schema documents and into separate parts of the back matter. The changes were integrated into the status quo documents on 20 January 2006.