When a method calls itself

This you can't do in RPG - have a method call itself. Sounds like something you would never use? Think again.

See this small console program that recurses through directories to get a list of folders and files. When the method finds a directory, it calls itself with the sub folder as a parameter whereupon the File part of the IF statement is executed.

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.IO;


namespace ConsoleApplication1
{
    class Program
    {
        static void Main(string[] args)
        {
            Recurse(@"c:\temp");
        }
       
               public static void Recurse(string directory)
    {
        DirectoryInfo path = new DirectoryInfo(directory);
        FileSystemInfo[] files = path.GetFileSystemInfos( );
        foreach (FileSystemInfo file in files)
        {
            if (file is DirectoryInfo)
            { 
                Console.WriteLine("Folder-> " + ((DirectoryInfo)file).FullName);

// Now the method calls itself passing in the subfolder name. When the method is called
all the files in the subfolder are listed 
                Recurse(((DirectoryInfo)file).FullName);
            }
            if (file is FileInfo)
            {
                Console.WriteLine("File-> " + ((FileInfo)file).FullName); 
            } 
            
        }
    


        }

    }
}

LiveJournal Tags: recursion,i/o,filesystem,directory,folders

C# -'The worst bloody language in the world'

My brother-in-law, a Phd wielding university prof.who has a penchant for the complex and abstruse called around yesterday to declare the whole .Net  bouquet and the languages contained therein 'the worst bloody language in the world'.  He came to me asking some very simple questions such as how to read a customer file, display it in a windows form, process it etc and was soon bogged down in connection strings, data binding, data sets etc. "But I just want to display the data and read it!" he exclaimed. He was astounded that the columns in a table aren't easily directly accessible when using SQL which isn't checked until run-time (LINQ I suggest? " Yeah but you still have to go thru hoops!"). 
Coming from procedural languages like RPG, I can sympathize. If I want to process a table in RPG  to check spending limits of a customer and update the table I do this

F CustomerFile IF   K DISK

C  Read CustomerFile;
C  DOW not %eof;
C  If AmountSpent > CreditLimit;
C  AllowSpending =False;
C  Update CustomerFileR;   
C  Endif;

C  Read CustomerFile;
C EndDO;

The equivalent in C# is much more complicated and includes connection strings, command text and a data reader which all have to be set up.

Note that for the ONE 'F' declaration in RPG there are 3 in C#.  To even do a read needs to be setup - you need to create a data reader first. Imagine having to 'create' a read statement in RPG! The code doesn't even include the update functionality!  But the biggest problem is that column names are not typed directly in C# meaning you can't refer to the column name directly. The column 'AmountSpent' is not known to the C# program.   Linq alleviates this situation but you still have to do the setup work first.  I understand the frustration.

using System;
using IBM.Data.DB2.iSeries;

namespace iSeriesADOexample
{
    class Program
    {
        static void Main(string[] args)
        {
            iDB2Connection connection = 
                new iDB2Connection("DataSource=PUB1.RZKH.DE; UserID=XXX; Password=xxx; DefaultCollection=COLMBYRNE1; LibraryList=COLMBYRNE1, *USRLIBL");
            iDB2Command cmd = connection.CreateCommand();
            connection.Open();
            
            cmd.CommandText = "Select * from COLMBYRNE1.CUSTOMERFILE";

            try
            {
                iDB2DataReader dataReader = cmd.ExecuteReader();
                while (dataReader.Read() == true)
                {

                   Double AmountSpent =  dataReader.GetDouble(3);
                    Double CreditLimit = dataReader.GetDouble(4);

                    if (AmountSpent > CreditLimit)
                    {
                        DoUpdate();
                    }

                        }
                }
                catch (iDB2SQLErrorException e)
            {Console.WriteLine("Error:" + e.MessageDetails);
                }
            Console.Read();
            cmd.Dispose();
        }
        }

        }
    

Programming is the art of bringing ideas to life. That's what my brother-in-law came to me for. He had an idea for a program and wanted to make it real. Unfortunately the initial  hurdle to create his simple program was far too great. The wizard functionality in Visual Studio only highlights how difficult it is to do basic actions like read a table, display the contents, process the results and update the table. This is one of the most common programs that every programmer creates yet in .NET it is difficult for a beginner.

It is clear Microsoft have 5-10 years to go before .NET is really mature and allows both beginner and seasoned developer to easily bring their ideas into the world. We program to make things better, faster, more fun, more interesting - not harder.  The tools we use then should also be better, faster - not harder. The effort to use a tool must never be greater than the effort it takes to solve the problem logically.  (Byrne's first hypothesis*) If you want to add 2 and 5 , C# should do that as easily as it does to solve it- and it does  int sum = 2+ 5;    E.g. if a client says that some of his customers are over their credit limit and needs to halt their spending then the logical solution which is to flag those spenders. The effort to implement that in C# should be that easy.  This is the benchmark which Microsoft must follow - just to keep up.

RPG Subroutine = C# Method?

Someone asked me the other day if RPG subroutines are the same as C# methods.
The answer is yes - like C# static methods but RPG Procedures are even more similar to C# static methods.
The reason being is that while C# methods are discrete named blocks of code you can call from within a program (C# Class) - just like an RPG Subroutine; C# methods allow parameters in the call - just like RPG procedures do. Methods have lots of other features that I’ll get in to but first lets draw comparisons.
Here’s an example:
Both applications read a table called ‘People’ and print out the name of each person on the console.
They both call a subroutine/method called Get_People that takes no parameters from the main block of code.

RPG
Fpeople IF E K DISK
/FREE
exsr Get_People ;
*inlr = *on;

BegSr Get_People;
read People ;
dow not %eof ;
Dsply name ;
Read People
Enddo;
Endsr;


C#

using System;
using System.IO;
namespace ConsoleApplication1
{
class ProgramClass
{
static void Main() // Main block of code 'Main' must be in the class
{
get_People();
}
static public void get_People( )
{
using (FileStream fileStream = new FileStream(@"C:\csharp\people.txt",
FileMode.Open,
FileAccess.Read,
FileShare.None))
{
using (StreamReader streamReader = new StreamReader(fileStream))
{
string text = streamReader.ReadLine();
while (text != null)
{
Console.WriteLine(text);
text = streamReader.ReadLine();
}
}
}
Console.ReadLine(); // Pause the screen
}
}
}

As you can see, RPG is easier to read and less verbose than its C# counterpart. To be fair DB2 is part of the iSeries operating system so all the file handling is already part of the DB.

Remember these are Static methods shown in C#. A static method in C# is where there is only one instance of the method. Believe it or not you can multiple instances of a method in C#. These are called instance methods and are used more often than static methods. I'll get into the difference but the purpose of this post is to show some similarity between RPG and C# first.

Comparing RPG code to C# : The For Loop

There are similarities between free format RPGIV and C#. In the example below I show a small program that accomplishes the same task in both languages. The task is to find the first blank character in a string from the end to the beginning. I'm well aware there is a single method available for this on String in C# but the purpose here is to show the similarities.
RPG is below:


















Now C#:

You can see some basic similarities:

• Statements end with a ';' ,


• the For loop is similar.


• 'Leave' gets you out of an RPG loop while 'Break' exits a C# loop.


• Assigning values is the same in both.


• Declaring variables in RPGIV requires both type and how many characters (you can declare varying length variables but do define an initial length). To define a variable you must put values in fixed positions in a special kind of statement block called 'Definition Specifications' or D-specs. In C# you simply declare the type. The type of the index 'i' is denoted by a zero in the decimal placess position.


• C# requires a 'Main' method. RPG just executes from the first executable line.


• C# variables are local by default. I can't access that 'i' variable outside of that For loop - the program wouldn't even compile. RPG variables are global unless used in procedures.



• C# requires a class definition, namespace and you define what other namespaces in .Net you want to refer to in your program with the 'Using' statements. The using just means you don't have to refer to the full name like 'System.Console.Readline()'.



• C# doesn't need 'EndFor' or 'EndIf'. Coding is accomplished in blocks between curly braces '{' and '}'.



• The biggest difference is that C# strings and objects have a multitude of methods and properties on the string itself - not part of the language. RPGIV is pretty good at string manipulation too but can never approach the flexibility of C# as the RPG designers would have to create a new keyword when it needs to perform a function on a string that it doesn't already have. In C# there are only a handful of keywords - most actions are accmplished by methods on the objects themselves. For instance in the above example you access the length of a string by examining its property. e.g. 'field.length' returns an integer value whereas in RPG I chose to use the built-in-function %len to do the same thing.
  To find the first non-blank in a string starting from the end you can use field.IndexofLast(' ') . RPGers note how the methods and properties are on the string 'field' which I defined. As soon as you define a variable it creates a copy of the string object so you have full access to a huge amount of manipulation techniques. Adding new methods and properties to strings and other types isn't a major change in the language.

RPG Output below

C# Output: