- From: C. M. Sperberg-McQueen <cmsmcq@acm.org>
- Date: Mon, 09 Oct 2000 21:16:23 -0600
- To: W3C XML Schema Comments list <www-xml-schema-comments@w3.org>
This note should have been copied to the XML Schema comments list. >Date: Mon, 09 Oct 2000 21:05:34 -0600 >To: Philip Wadler <wadler@research.bell-labs.com>, Paul Cotton ><pcotton@microsoft.com>, W3C XML Query WG <w3c-xml-query-wg@w3.org> >From: "C. M. Sperberg-McQueen" <cmsmcq@acm.org> >Subject: LC-166 alignment of simple and complex types > >Dear Philip, Paul, and members of the XML Query WG: > >The W3C XML Schema Working Group has spent the last several months >working through the comments received from the public on the last-call >draft of the XML Schema specification. We thank you for the comments >you made on our specification during our last-call comment period, and >want to make sure you know that all comments received during the >last-call comment period have been recorded in our last-call issues >list (http://www.w3.org/2000/05/12-xmlschema-lcissues). > >Among other issues, you raised (separately) the point registered as >issue LC-166, which suggests that there is a lack of orthogonality in >XML Schema in that simple and complex types may not always be used in >the same way, and proposes that the names of simple types be allowed >wherever the names of complex types can be used. > >We agree that there should be orthogonality in this area, but are >puzzled by your claim that there is none. With two exceptions, the >names of simple and of complex types can in fact be used in the same >places. > > (1) Only simple types are allowed as the types of attributes. This > is a consequence of the basic design of SGML and XML; it is not > feasible for XML Schema to change it. > > (2) Simple types may not be derived from complex types, and thus the > base type definition of a simple type must be a simple type and not > a complex type. > >We believe, that is, that with respect to the definition of complex >types (which is the context you appear to have in mind), the state of >affairs you describe as your goal is already the case. > >Your messages suggest a number of consequences which would, you >suggest, follow from the change you request. It is the belief of some >members (at least) of the XML Schema WG that this list of consequences >would follow not from allowing simple types to be named wherever >complex types may be named, but from allowing them to be named as >particles within complex types in the same way that elements may be >named; since this is not currently legal for complex types, this would >mean introducing, not eliminating, a lack of orthogonality between >simple and complex types. > >The proposal to allow simple types to occur as particles within >complex types (so that in an element like <frac>143<bar/>798</frac> >the "143" and the "798" could each be typed as integers) is similar to >the proposal (in issue LC-51) to introduce an analog to the #PCDATA >particle in SGML element type declarations. The WG declined to make >that change, partly because it would make content models more complex, >and partly because it would reintroduce some form or other of the >extremely vexing complications for parsers which are well known in >SGML (usually referred to as "the mixed-content gotcha") and have been >mercifully eliminated from XML. The addition of type information to >the question (i.e. using references to simple types instead of a >PCDATA particle) further increases the complexity, and does not make >the proposal any more attractive. > >The most serious reason to decline the proposal, however, is this: The >simple fact that in some places within the element, character data is >allowed, and in other cases it is forbidden, suggests that there is >some semantic difference between those two regions. But it is an >inherently questionable design (if not necessarily always a wrong >design) to identify an important semantic unit and then to define no >element type in the markup language corresponding to that semantic >unit. The WG did not feel any pressing urge to make it easy to >implement, in a schema, such questionable design decisions. > >The fact that the change you propose would introduce an unmotivated >difference between the treatment of simple and complex types is also >an argument against it which we expect you will appreciate. > >It would be helpful to us to know whether you are satisfied with the >decision taken by the WG on this issue, or wish your dissent from the >WG's decision to be recorded for consideration by the Director of the >W3C. > >with best regards, > >-C. M. Sperberg-McQueen > World Wide Web Consortium > Co-chair, W3C XML Schema WG
Received on Monday, 9 October 2000 17:18:37 UTC