- From: Joseph Kesselman/Watson/IBM <keshlam@us.ibm.com>
- Date: Tue, 19 Sep 2000 08:58:35 -0400
- To: www-dom@w3c.org
I'm not sure I'm following your question. Node.nodeValue may be either an empty string or null, depending on the node in question. For example, Element and Document nodes have no nodeValue, and return null. Text nodes may be empty, and hence may return an empty string. Similarly, a null Namespace URI is significantly different from a blank Namespace URI -- the latter is namespace-ill-formed, the former is just a node which has not been assigned to a namespace. If the language/binding conventions that you're working with don't intrinsically support null, it's the responsibility of whoever designs that set of bindings to come up with an equivalent. Whether that's a specific string value (presumably one that would otherwise be illlegal in XML), or a specific string object (if your system lets you compare objects for identity and treats strings as objects) seems to be their choice. I don't know a great deal about CORBA, but if sequence<unsigned short> doesn't support null, maybe it simply isn't the best datatype to directly bind as DOMString. ______________________________________ Joe Kesselman / IBM Research
Received on Tuesday, 19 September 2000 08:58:39 UTC