- From: Prasad Yendluri <pyendluri@webmethods.com>
- Date: Thu, 27 Jun 2002 16:17:14 -0700
- To: David Cleary <davec@progress.com>
- CC: www-ws-desc@w3.org
David, Taking a careful look I realized only SOAP and HTTP bindings schemas do not have local declamations of elements. Both WSDL and MIME binding schemas do have local element declarations (especially WSDL one has numerous). So you were mistaken on your assertion below: > BTW, the WSDL schema doesn't appear to use local element definitions in any > case, making the argument moot. BTW, are you the same David Cleary that used to work for webMethods :)? Regards, Prasad Prasad Yendluri wrote: > David, > > I was not suggesting we change the spec based on the examples. I was citing > examples to support my issue. Anyhow I agree with your other points.. > > Regards, Prasad > > David Cleary wrote: > > > > The WSDL Schema ([1] or [2]) sets the elementFormDefault="qualified". > > > This, AFAIK requires each element to qualified in the instance > > > documents, requiring one to use ns qualifiers with many of the elements > > > defined in the WSDL spec (<wsdl:message ..> <wsdl:service ..> > > > <wsdl:portType ..> etc. where the namespace > > > wsdl="http://schemas.xmlsoap.org/wsdl/"). Most of the WSDL instance > > > examples in the spec violate this. Why do we need to keep this? Can we > > > change elementFormDefault="unqualified" and be done with it? Or am I > > > mistaken here? > > > > Many of the examples in the spec suffer from errors, but they should be > > fixed instead of changing the spec. Unless you have a reason why local > > elements should not be qualified, the status quo should be kept. A WSDL > > document contains multiple namespaces. Arbitrarly changing the schema to use > > unqualified local elements without a good reason will make reading and > > processing WSDL more error prone and harder to read. Unqualified elements > > should be used in situations similar to the use of unqualified attributes. > > If the data contained in the element is useless without its parent or > > siblings, using unqualified elements to reinforce this fact makes sense. > > However, that is not the case here. > > > > BTW, the WSDL schema doesn't appear to use local element definitions in any > > case, making the argument moot. > > > > David Cleary
Received on Thursday, 27 June 2002 19:13:36 UTC