A typical IRS Job has a number of steps
        for example
        
        Set #1 - ALC
            #2 - COBOL
            #3 - AlC
            etc.
            
        The IRS is moving away from ALC and COBOL
        this means they are converting a single step with a job at a time
        this mean the new Java appl to replace the ACL or COBOL step
        MUST read/write files create/consumed by ACL or COBOL
        
        for example the previous COBOL step wrote this binary file:
        

      * COPYBOOK FOR THE VEHICLE ORDER RECORDS
       01  Order.
           05 ID                    PIC 9(04).
           05 DEALER-ID             PIC 9(06) COMP.  <- this is a BINARY number value
           05 ORDER-DATE.
              10 ORDER-DAY          PIC 9(02).
              10 ORDER-MONTH        PIC 9(02).
              10 ORDER-YEAR         PIC 9(04).
           05 MAKE                  PIC X(20).
           05 MODEL                 PIC X(25).
           05 VIN                   PIC X(17).
           05 YEAR                  PIC 9(04).
           05 BASE-COST             PIC S9(07)V99 COMP-3. <- this a a PACKED-DECIMAL number value
           05 TAX-RATE              PIC 9V99.             <- this has an implied decimal place
           05 TOTAL-COST            PIC S9(07)V99 COMP-3.
           
           
      so...
      
      My Java code MUST be able to deal with these file formats
      
      ibmjzos.jar to the rescue
        
        
     the steps to access a COBOL dataset from Java are:
     
     1) get a copy of the COBOL Copybook that defines the file layout
     
     2) I write a Java Class to act as a converter, I will have methods to
         get/set each field in the file
         This is the ONLY place the Java code deals with the feilds
         
         In this class I create a IBM Datatype field for EACH COBOL field

    //  05 ID                    PIC 9(04).  <- the COBOL field as a comment
        private static final ExternalDecimalAsIntField id =      <- the IBM Datatype
                        factory.getExternalDecimalAsIntField(4, false);  <- invoke the factory
    //  05 DEALER-ID             PIC 9(06) COMP.
        private static final BinaryAsIntField dealerId = 
                        factory.getBinaryAsIntField(6, false);   
                        
        after I define the fields I add get/set method for the Java code to call
        
        public int getId() 
        {
                return id.getInt(bytes);  <- note the COBOL file is binary so I need an array
        }                                      of bytes

        public void setId(int idValue)    <- the get/set method deal ONLY w/ Java datatypes
        {
                id.putInt(idValue, bytes);
        }   
        
        
     3) In my Java code I call the converter to read a line from the Cobol file
         and pass the byte[] to the converter
         the call the get methods to extract teh Java values fro the Cobol filed
         
         or
         
         I create a empty converter to write out a Cobol file, I call teh set
         methods to load the fields from the Java fields and get the byte[] 
         I can use to write out a binaly line into the new Conol file