- From: <keshlam@us.ibm.com>
- Date: Tue, 9 Nov 1999 09:12:28 -0500
- To: www-dom@w3.org
The problem here may be multiple definitions of the word "package", and multiple packages in a single application... Let's try a concrete example: IBM's XML4J parser, which currently includes two implementations of the DOM. The DOM API Interfaces themselves (Node.java and so on) are included un-altered as the org.w3c.dom package. The implementations live in their own packages (eg com.ibm.xml.dom). They reference the DOM APIs, and are operated primarily in terms of the DOM APIs, but their code is kept distinct from the W3C's code. They may provide additional functions, or implement additional interfaces, as well. All these packages wind up in a single jar-file when IBM distributes the compiler. But by keeping the packages distinct, they maintain a clear line between which features are defined by the DOM and which are custom additions. Doing so ensures that if this code is linked with other DOM-based code, compiled seperately with another copy of org.w3c.dom, everything will connect up properly. Even without the copyright issue, that is a Darned Good Reason to do it this way. ______________________________________ Joe Kesselman / IBM Research
Received on Tuesday, 9 November 1999 09:12:51 UTC