Based on some interest expressed in having Apache SOAP support the Oracle convention of having Node.namespaceURI return an empty string instead of null when there is no namespace, I posted to [EMAIL PROTECTED] about whether the spec should further clarify what is meant by "null". This is the only response to date.
Scott Nichol ----- Original Message ----- From: "Fred L. Drake, Jr." <[EMAIL PROTECTED]> To: "Scott Nichol" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Friday, July 25, 2003 11:37 AM Subject: Re: Node.namespaceURI > > Scott Nichol writes: > > This regards Node.namespaceURI in the DOM 3 Core. Going back to > > DOM 2, this attribute's description has started "The namespace URI > > of this node, or null if it is unspecified". To me, this seems > > quite clear, and in programming languages like Java, I expect a > > null value in that language to be returned. > > For Python, null is mapped to the None value, which is generally used > in Python to indicate "no value". The Python XML-SIG has gone round > and round on the issue of whether to use an empty string or None for a > missing namespace URI in the DOM, and we finally settled on None since > that's really what's intended. > > > However, there are implementations that do not use what I would > > consider to be a null. For example, Oracle's XML parser's > > Node.getNamespaceURI() returns an empty (zero-length) string, and > > Microsoft's .NET framework (the NamespaceURI property of the > > System.Xml.XmlNode class) likewise returns an empty string. > > This sounds like a serious nuissance in those implementations. > > But I think the spec spells out clearly that null is the right thing. > If that's normally spelled using Java's null value, there's no need to > change the specification; the implementations are clearly in violation > and bug reports should be filed. > > > -Fred > > -- > Fred L. Drake, Jr. <fdrake at acm.org> > PythonLabs at Zope Corporation
