- From: Jacek Kopecky <jacek@systinet.com>
- Date: Wed, 3 Jul 2002 18:59:58 +0200 (CEST)
- To: Kevin Johnsrude <kevinj@roguewave.com>
- cc: xml-dist-app@w3.org
Kevin, first, I apologize for taking so long to write this. It is true that SOAP Encoding document is not fully self-descriptive. It may not be a problem, though, see below. First, let me summarize the possible ambiguities: 1) There is no difference between an empty simple value and a compound value with no members. 2) One cannot tell whether a struct is not, in fact, an array or a generic. 3) One cannot tell whether an array without array metainformation is not, in fact, a generic, or in some cases a struct. 4) In some cases, one cannot tell whether a generic is not, in fact, a struct. 5) One cannot tell whether a generic is not, in fact, an array. This is a problem unless we expect the receiver to know the structure of the data (to have some form of schema for the data). SOAP 1.2 doesn't provide any formal schema language for its data model because it didn't seem necessary (the applications will somehow manage). It is my understanding that SOAP expects other parties to provide the information necessary for understanding SOAP Encoding data instances. For example, WSDL 1.1 has the so-called "encoded" use of XML Schema schemas that is used exactly for this. Personally, I think that the party providing the data model should also provide a way to specify schemata in that data model, but if the rest of the world thinks otherwise, I won't stand in the way. 8-) Best regards, Jacek Kopecky Senior Architect, Systinet Corporation http://www.systinet.com/ On Fri, 28 Jun 2002, Kevin Johnsrude wrote: > > In SOAP 1.2, 3.1.2 "Encoding compound values" [1], is there any way to > distinguish between a "struct" and an "array" that has neither an "itemType" > nor an "arraySize" attribute? Note that this appears to be permitted per > item 3. > > [1] http://www.w3.org/TR/2002/WD-soap12-part2-20020626/#complexenc > > Cheers, > Kevin Johnsrude > Rogue Wave Software, www.roguewave.com >
Received on Wednesday, 3 July 2002 13:00:00 UTC