/*
 * @(#)MultiDoc.java	1.3 03/01/23
 *
 * Copyright 2003 Sun Microsystems, Inc. All rights reserved.
 * SUN PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
 */

package javax.print;

import java.io.IOException;

/**
 * Interface MultiDoc specifies the interface for an object that supplies more 
 * than one piece of print data for a Print Job. "Doc" is a short, 
 * easy-to-pronounce term that means "a piece of print data," and a "multidoc" 
 * is a group of several docs. The client passes to the Print Job an object
 * that implements interface MultiDoc, and the Print Job calls methods on
 *  that object to obtain the print data.
 * <P>
 * Interface MultiDoc provides an abstraction similar to a "linked list" of 
 * docs. A multidoc object is like a node in the linked list, containing the 
 * current doc in the list and a pointer to the next node (multidoc) in the 
 * list. The Print Job can call the multidoc's {@link #getDoc() 
 * <CODE>getDoc()</CODE>} method to get the current doc. When it's ready to go 
 * on to the next doc, the Print Job can call the multidoc's {@link #next() 
 * <CODE>next()</CODE>} method to get the next multidoc, which contains the  
 * next doc. So Print Job code for accessing a multidoc might look like this:
 * <PRE>
 *      void processMultiDoc(MultiDoc theMultiDoc) {
 *          
 *          MultiDoc current = theMultiDoc;

 *          while (current != null) {
 *              processDoc (current.getDoc());
 *              current = current.next();
 *          }
 *      }
 * </PRE>
 * <P>
 * Of course, interface MultiDoc can be implemented in any way that fulfills  
 * the contract; it doesn't have to use a linked list in the implementation. 
 * <P>
 * To get all the print data for a multidoc print job, a Print Service 
 * proxy could use either of two patterns: 
 * <OL TYPE=1>
 * <LI>
 * The <B>interleaved</B> pattern: Get the doc from the current multidoc. Get 
 * the print data representation object from the current doc. Get all the print 
 * data from the print data representation object. Get the next multidoc from 
 * the current multidoc, and repeat until there are no more. (The code example 
 * above uses the interleaved pattern.) 
 * <P>
 * <LI>
 * The <B>all-at-once</B> pattern: Get the doc from the current multidoc, and 
 * save the doc in a list. Get the next multidoc from the current multidoc, and 
 * repeat until there are no more. Then iterate over the list of saved docs. Get 
 * the print data representation object from the current doc. Get all the print 
 * data from the print data representation object. Go to the next doc in the 
 * list, and repeat until there are no more. 
 * </OL>
 * Now, consider a printing client that is generating print data on the fly and 
 * does not have the resources to store more than one piece of print data at a 
 * time. If the print service proxy used the all-at-once pattern to get the 
 * print data, it would pose a problem for such a client; the client would have 
 * to keep all the docs' print data around until the print service proxy comes 
 * back and asks for them, which the client is not able to do. To work with such 
 * a client, the print service proxy must use the interleaved pattern. 
 * <P>
 * To address this problem, and to simplify the design of clients providing 
* multiple docs to a Print Job, every Print Service proxy that supports 
 * multidoc print jobs is required to access a MultiDoc object using the 
 * interleaved pattern. That is, given a MultiDoc object, the print service 
 * proxy will call {@link #getDoc() <CODE>getDoc()</CODE>} one or more times 
 * until it successfully obtains the current Doc object. The print service proxy 
 * will then obtain the current doc's print data, not proceeding until all the 
 * print data is obtained or an unrecoverable error occurs. If it is able to 
 * continue, the print service proxy will then call {@link #next() 
 * <CODE>next()</CODE>} one or more times until it successfully obtains either 
 * the next MultiDoc object or an indication that there are no more. An 
 * implementation of interface MultiDoc can assume the print service proxy will 
 * follow this interleaved pattern; for any other pattern of usage, the MultiDoc 
 * implementation's behavior is unspecified. 
 * <P>
 * There is no restriction on the number of client threads that may be 
 * simultaneously accessing the same multidoc. Therefore, all implementations of 
 * interface MultiDoc must be designed to be multiple thread safe. In fact, a 
 * client thread could be adding docs to the end of the (conceptual) list while 
 * a Print Job thread is simultaneously obtaining docs from the beginning of the 
 * list; provided the multidoc object synchronizes the threads properly, the two 
 * threads will not interfere with each other
 */

public interface MultiDoc {


    /**
     * Obtain the current doc object.
     *
     * @return  Current doc object.
     *
     * @exception  IOException
     *     Thrown if a error ocurred reading the document.
     */
    public Doc getDoc() throws IOException;

    /**
     * Go to the multidoc object that contains the next doc object in the
     * sequence of doc objects.
     *
     * @return  Multidoc object containing the next doc object, or null if
     * there are no further doc objects.
     *
     * @exception  IOException
     *     Thrown if an error occurred locating the next document
     */
    public MultiDoc next() throws IOException;

}