This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
QT approved comment: In 3.1, in the sentence "Each user-defined datatype is also associated with a unique namespace. ", it is not clear what "unique" adds to the sense. [It appears to be saying that two datatypes cannot share the same namespace. That's obviously a misreading, so what *is* it saying?] It would also be nice to have a name for this "association": the following sentence calls the relationship "comes from" which seems slightly bizarre, and is not used consistently elsewhere. "Belongs to" might be better.
Agreed that the sentence "Each ·user-defined· datatype is also associated with a unique namespace" is problematic. Not only does it invite the reader to wonder what the word 'unique' is doing loitering in this neighborhood, it also suggests falsely that user-defined datatypes always have a target namespace. (Life would perhaps be simpler if unqualified names were in a sort of special anonymous default namespace, but they are not. Complaints to the editors of the Namespaces Rec, please.) Agreed also that the following sentence is awkward. I propose the following revision. Delete the paragraph which currently reads: Each ·user-defined· datatype is also associated with a unique namespace. However, ·user-defined· datatypes do not come from the namespace defined by this specification; rather, they come from the namespace of the schema in which they are defined (see XML Representation of Schemas in [XML Schema Part 1: Structures]). Replace it with: Each ·user-defined· datatype MAY also be associated with a namespace. However, ·user-defined· datatypes are not defined in the XML Schema namespace (http://www.w3.org/2001/XMLSchema); rather, they are associated with the target namespace of the schema in which they are defined (see XML Representation of Schemas in [XML Schema Part 1: Structures]). Implicitly, this wording proposes the phrase "is defined in" (or just "is in", except that many readers will want "is in" to apply only to names, not type definitions) or "is associated with" for the relation of a simple type definition or datatype to its target namespace. I'm setting the status of this issue to needsReview, but it should be noted that this wording proposal has NOT had the normal editorial review; the other editors should thus not be held accountable for it.
Looks good. >(Life would perhaps be simpler if unqualified names were in a sort of special anonymous default namespace, but they are not. Complaints to the editors of the Namespaces Rec, please.) I did complain to them and it did no good. Perhaps that's because I complained about nearly everything in their spec. But we could make life simpler for ourselves. Just define that "in this specification", namespace means either (a) a namespace as defined in the namespaces Rec, or (b) the set of names which, in the terminology of the namespaces Rec, are not in any namespace; this specification refers to this set of names as "the unnamed namespace".
proposal withdrawn this date for further revision
As a result of a recent change to the Datatypes spec (editorial proposal 26, to be specific, as amended by the WG), the paragraph in question no longer contains the word "unique". The paragraph now reads: Each ·user-defined· datatype may also be associated with a target namespace. If it is constructed from a schema document, then its namespace is typically the target namespace of that schema document. (See XML Representation of Schemas in [XSD 1.1 Part 1: Structures].) The main part of the issue has thus been addressed. The other part (having a better name for the association of datatypes with namespaces) has not been addressed, and seems apt to lead to a long battle over terminology. I propose (this has not been discussed among the editors, and does not necessarily represent editorial consensus) that we close this issue with the keyword 'FIXED', and ask QT to accept this resolution.
The XML Schema Working Group discussed this issue today and decided to adopt the proposal in comment #4, namely to close this issue as having been resolved (at least for the most part) by changes made for other issues. The suggestion that a better term for the relation between names and their namespaces is not unwelcome, but we have not found a really satisfactory name, so we do not expect to make textual changes for the purpose of changing that terminology. Michael Kay, as the originator of the issue, would you report back to QT on this resolution and let us know whether they accept this resolution of the issue? If they agree, please so indicate by changing the record's status to CLOSED; if they disagree, REOPEN it. If we don't hear from you in a reasonable amound of time (say, two weeks), we will assume that silence implies consent.