W3C home > Mailing lists > Public > xml-dist-app@w3.org > May 2002

Re: Rewording of section 4.1.2 based upon resolution of issue 195

From: Ray Whitmer <rayw@netscape.com>
Date: Thu, 16 May 2002 11:26:02 -0700
Message-ID: <3CE3F9BA.7010509@netscape.com>
To: Marc Hadley <marc.hadley@sun.com>, xml-dist-app@w3.org
Marc Hadley wrote:

>> An RPC response is modeled as a struct where parameter access is by 
>> name or as an array where parameter access is by position. [The  SOAP 
>> encoding specification defines no way to directly determine whether 
>> the response is modeled as a struct or as an array.]
> Doesn't examination of the response do that: itemType present - array, 
> arraySize present - array, all of the child elements the same name - 
> array otherwise a struct or (perhaps) a generic compound type ?

I believe that all elements of an array are not required to have the 
same name.  While it is a possible indicator, it does not provide a sure 
determination whether the responseis an array or struct.  Neither 
arraySize nor itemType are mandatory, so as with elements having the 
same name, it may or may not be true.

An implementation might produce a call or reponse that could legally be 
interpreted as either a struct and array, depending upon which schema 
was used to interpret it.  As long as it didn't specify an arraySize or 
array itemType, and all elements were differently-named, this would seem 
to be no problem except that a return value would stretch it up a bit 
since it adds an edge in the struct case, but it wouldn't be too hard to 
put that in an ignorable position.

Best regards,

Ray Whitmer
Received on Thursday, 16 May 2002 14:24:51 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:01:20 UTC