- From: Hank Davidson <hankd@corel.com>
- Date: Mon, 18 May 1998 10:05:36 -0600
- To: "Mike Champion" <mcc@arbortext.com>, <www-dom@w3.org>
A couple more questions, please. 1. What happens to the node that is passed in to the Attribute factory; does it become the child of the attribute? 2. If so, how do you construct the attribute when it may have multiple children, e.g. text - entity reference - text? 3. If not, what type of node is it that is passed in? Do its children become the children of the resulting attribute node? Does this mean that to set a simple text value, it would be represented by a two-deep hierarchy as the parameter to the constructor? -----Original Message----- From: Mike Champion <mcc@arbortext.com> To: www-dom@w3.org <www-dom@w3.org> Date: Friday, May 15, 1998 2:27 PM Subject: Re: Why is the value of an Attribute a Node? >At 03:55 PM 5/15/98 -0400, Hank Davidson wrote: >>The java interface for Document specifies that the arguments to the >>createAttribute method are the name (a String) and a value (a Node). >>Why is value a Node? It seems inconsistent with the getValue method on >>Attribute which returns a String, and the setAttribute method on Node >>which takes a String for the value. > >I think there's some errors in the current draft of the spec (the Attribute >stuff was discussed at the last minute before the draft went out), and >things DEFINITELY should be explained more clearly. > >Conceptually, Attribute values are Nodes because in XML, entity references >in can expand to make the value an arbitrarily complex tree, and the Node >is the root of that tree. In HTML, and in probably the vast majority of >XML applications, the value is a string. SO, we formally define it as a >Node, but give convenience methods to get/set attribute values as strings >since that will be the most common usage. > >I'll clarify all this in the next draft. > >Mike Champion > >
Received on Monday, 18 May 1998 12:05:38 UTC