5637 2008-04-10 16:25:08 +0000 inconsistencies in language for rule attachment to schema components 2008-12-09 18:02:47 +0000 1 1 1 Unclassified SML Core LC PC Windows XP RESOLVED WORKSFORME P2 normal --- 5636 1 johnarwe cmsmcq public-sml oldest_to_newest 19769 0 johnarwe 2008-04-10 16:25:08 +0000 As I read through 6.3 Rules Associated with Schema Components with my "unfamiliar outsider" hat on, I worry about the apparent (at least - might be real problem) inconsistencies about where the spec uses/omits the qualifier "global" when talking about elements and types. 6.3 defines {rules} for all (not just global) element decls and type defs. 6.3.1 says - sch:schema elements may be embedded ONLY in GLOBAL EDs/TDs. - local-rules defined only for GED/GTD - defines the value of {rules} for xs:anyType (not a GTD AFAIK) - bullet 3 omits global 6.3.2 says - bullet 1 echo's 6.3.1 1st "-" point above, i.e. ONLY GLOBALS It also specifically mentions anonymous types of GEDs, which seems "odd" given that they are mentioned nowhere else that I can find. - bullet 2 omits global from type decl, includes it for element decl 6.3.3 says - bullet 1 omits global I think the wg has been clear enough in previous discussions that we understand the intent to be "only on GEDs and GTDs". This is simply a question of whether or not the current text clearly and accurately communicates that intent. 19776 1 virginia.smith 2008-04-10 19:05:07 +0000 Discussed in 4/10 call: MSM: xs:anyType is in fact a GTD MSM: if we make any changes based on 5636, may need to hit the text for this bug. 20846 2 johnarwe 2008-06-25 11:55:58 +0000 f2f consensus while some members feel that naive readers might be confused by the apparent inconsistencies, the wg is confident that the existing language is correct, and no one is able to make a concrete proposal to reduce such confusion.