ZH's software thread

[funkyass]

Well, I'm not saying that indenting isn't a good thing. I just don't like being forced to do it.

you'll love perl then.

[Francis64]

Code: Select all

    int lowest = 0;
    int highest = 0;
    while ( true ) {
          cout << "Enter number (0 to finish): ";
          cin >> num;
          if ( num == 0 ) {
               return 0;
          }

          if ( num < lowest ) {
              lowest = num;
          } else if ( num > highest ) {
              highest = num;
          }

          cout<<"Lowest number "<< lowest <<" and highest number "<< highest << endl;
    }

Simplified for you. You're doing more tests than necessary.
And well done on your first C++ program[me].

[creaothceann]

Well, I'm not saying that indenting isn't a good thing. I just don't like being forced to do it.

There are a lot of unwritten rules. Goto is bad, separate control code from implementation code, write small functions that do one thing, indendation is good, consistent indendation is better, etc.

Unless you know several languages, and know when to temporarily ignore the rules, you have no authority to bitch about them.

[byuu]

Goto is bad

Oh yeah, definitely.

Code: Select all

bool epicfail() {
  bool success = false;

  resource *a = open_file_source();
  if(!a) goto This;

  resource *b = open_file_dest();
  if(!b) goto Code;

  resource *c = allocate_memory();
  if(!c) goto Is;

  resource *d = get_exclusive_lock();
  if(!d) goto Bad;

  ...
  success = true;

  release_exclusive_lock();

  Bad:
  free_memory();

  Is:
  close_file_dest();

  Code:
  close_file_source();

  This:
  return success;
}

What kind of jackass writes code like that?

This is so much better:

Code: Select all

bool make_java_proud() {
	resource *a = open_file_source();
	if(a)
	{
		resource *b = open_file_dest();
		if(b)
		{
			resource *c = allocate_memory();
			if(c)
			{
				resource *d = get_exclusive_lock();
				if(d)
				{
					...
					release_exclusive_lock();
					free_memory();
					close_file_dest();
					close_file_source();
					return true;
				}
				else
				{
					free_memory();
					close_file_dest();
					close_file_source();
					return false;
				}
			}
			else
			{
				close_file_dest();
				close_file_source();
				return false;
			}
		}
		else
		{
			close_file_source();
			return false;
		}
	}
	else
	{
		return false;
	}
}

:D

[creaothceann]

Ah, excellent case for exceptions!

Code: Select all

function MakeJavaProud : Boolean;
begin
Result := False;
if (not OpenFileSource(a)) then Exit;
try
        if (not OpenFileDest(b)) then Exit;
        try
                if (not AllocateMemory(c)) then Exit;
                try
                        if (not GetExclusiveLock(d)) then Exit;
                        try
                                //...
                                Result := True;
                        finally
                                ReleaseExclusiveLock(d);
                        end;
                finally
                        FreeMemory(c);
                end;
        finally
                CloseFileDest(b);
        end;
finally
        CloseFileSource(a);
end;
end;

Exeptions just look complicated (well, they are), unless you do something like this:

Code: Select all

function MakeJavaProud : Boolean;
begin
Result := False;
if (not OpenFileSource  (a)) then Exit;  try
if (not OpenFileDest    (b)) then Exit;  try
if (not AllocateMemory  (c)) then Exit;  try
if (not GetExclusiveLock(d)) then Exit;  try
        //...
        Result := True;
finally  ReleaseExclusiveLock(d);  end;
finally  FreeMemory          (c);  end;
finally  CloseFileDest       (b);  end;
finally  CloseFileSource     (a);  end;
end;

... which is not much easier to understand.

[ZH/Franky]

Well, I've found something that (for now) makes Python better than C++.
Compare the two images of the same program, but written in different languages (python and C++). Click the thumbnails to see them fullsize:
C++ version:

Python version:

they are both of the program I posted that find the factorial of a number. One is in C++ (the one I originally posted), and the other is a version of it in python. As you can see, python displays large numbers, where are C++ (or more specifically, double variables) resort to standard form. Standard form is cool, but sometimes it displays a "rounded up" (i,e, completely innacurate) number (standard form can't do everything, you know).
So, for large numbers, python is currently best. Still, right now I'm favouring C++ much more than I favour python.

[odditude]

double is by definition a floating point number stored as an exponential. you'd be better off using a long long (or whatever the proper terminology is now) if you want an integral result.

[grinvader]

Goto is bad

Oh yeah, definitely.

<epicfail()>

What kind of jackass writes code like that?

This is so much better:

<make_java_proud()>

:D

Yeah, you jackass. fclose and free and others don't die horribly if you pass them a null pointer, duh. And for those who do, wrap them in something that doesn't.
Forgot the obligatory ":D", for those who need it.

Code: Select all

bool use_brain()
{
  bool ret = -1;
  FILE *sauce_fp = fopen(input.fname, "rb");
  FILE *dest_fp = fopen(output.fname, "rb");
  char *buffer = (char *)malloc(BUFSIZE);
// ...

  if (sauce_fp && dest_fp && buffer /* && ... */)
  {
// stuff
    ret = 0;
  }

// ...
  free(buffer);
  fclose(dest_fp);
  fclose(sauce_fp);

  return (ret);
}

And if you CANNOT do something like this, then you should stop and try another design where you can.
Unless you like goto, of course. This coming from a consummate ASM whore.

[byuu]

fclose and free and others don't die horribly if you pass them a null pointer, duh.

Way to completely miss the whole god damn point by nitpicking :P

[ZH/Franky]

Ok, so anyway I did that bin2dec and dec2bin program, in visual basic:
Decimal to Binary:

Code: Select all

Module Module1
    Sub Main()
        Dim lp As Double
        Dim digit As Double
        Dim mem As Double
        Dim binr As String
        Dim BinaryString As String
        Dim pos As Double
        Dim master As Double
        master = 1
        Do While master = 1
            digit = 0
            binr = ""
            BinaryString = ""
            pos = 0
            Console.WriteLine("Enter a number to convert to binary: ")
            digit = Console.ReadLine()
            lp = 1
            Do While lp = 1
                digit = digit / 2
                mem = digit
                digit = digit - Int(digit)
                If digit > 0 Then
                    binr = binr + "1"
                End If
                If digit = 0 Then
                    binr = binr + "0"
                End If
                digit = mem
                digit = Int(digit)
                If digit = 0 Then
                    lp = 0
                End If
            Loop
            pos = binr.Length
            lp = 2
            Do While lp = 2
                BinaryString = BinaryString + Mid(binr, pos, 1)
                pos = pos - 1
                If pos = 0 Then
                    lp = 0
                End If
            Loop
            Console.WriteLine(BinaryString)
        Loop
    End Sub

End Module

Binary to decimal:

Code: Select all

Module Module1
    Sub Main()
        Dim pos As Double
        Dim BinaryString As String
        Dim binr As String
        Dim lp As Double
        Dim dec As Double
        Dim master As Double
        Dim negative As Double
        master = 1
        Do While master = 1
            pos = 0
            BinaryString = ""
            binr = ""
            dec = 0
            Console.WriteLine("Enter bin: ")
            BinaryString = Console.ReadLine()
            lp = 1
            pos = BinaryString.Length
            Do While lp = 1
                binr = binr + Mid(BinaryString, pos, 1)
                pos = pos - 1
                If pos = 0 Then
                    lp = 0
                End If
            Loop
            pos = 1
            lp = 1
            Do While lp = 1
                If Mid(binr, pos, 1) = "1" Then
                    dec = dec + (2 ^ (pos - 1))
                End If
                If Mid(binr, pos, 1) = "0" Then
                    dec = dec
                End If
                pos = pos + 1
                If pos > binr.Length Then
                    lp = 0
                End If
            Loop
            Console.WriteLine(dec)
        Loop
    End Sub

End Module

So what do you guys think?
I'm going to be writing a few more programs in vb soon:
- I'll improve upon the dec2bin program I posted here, so that it can convert negative decimal numbers into binary (that is, two's compliment)... I'll just make it so that when the user enters a minus number, it multiplies that number by minus 1, then convert the positive number it creates into a binary number, then change all the zero's to 1's, and all the one's to 0's, then convert that "flipped" binary number back into decimal, add 1 to that decimal, then convert that decimal back into binary. Easy stuff really. I just need to actually implement it in code
- I currently don't have an idea as to how to do this, but I'll write something that converts a two's compliment binary string to decimal (the user specifies that it's two's complement by prefixing the binary string they enter with a "t"; again, I'll implement this into the program that converts a postive binary number to a decimal
- when I've done both of the above, I'll merge both programs together, and write some code to give the user a little text menu so they can select what they want to do (bin2dec or dec2bin).
- After doing all three of the above, I will implement into this one program, the following things: to ability to do binary addition, multiplication, subtraction, and division.
- After doing all of this, I will try to write this whole program again, but in C++.

As of now, I mother fucking love programming.

[odditude]

these comments are on the top program.

the variable "master" serves no purpose. either give it a purpose or just nuke it. (do... instead of do while...)

instead of the bizarreness of the original being a double and you casting it as an int to get integer division, use the mod function.

actually, you could condense it all down to:

Code: Select all

...
dim digit as integer
...
Do
	binr = ""
	BinaryString = ""
	Console.WriteLine("Enter a number to convert to binary: ")
	digit = Console.ReadLine()
	Do While digit > 1
		binr = binr + (digit % 2)
		digit = digit / 2
	Loop
	binr = binr + digit
	pos = binr.Length
	Do While pos > 0
		BinaryString = BinaryString + Mid(binr, pos, 1)
		pos = pos - 1
	Loop
	Console.WriteLine(BinaryString)
Loop

also, if memory serves me correctly, there's a builtin to insert a character at the beginning of a string - that would negate the need for the flipper loop at the end.

edit: and ALWAYS validate your input.

[h4tred]

As of now, I mother fucking love programming.

Wait till you learn ASM. Then you will adore it.

[h4tred]

I will try to write this whole program again, but in C++.

this time, make a proper Win32 based GUI. >.<

[ZH/Franky]

edit: and ALWAYS validate your input.

I'm not trying to be a nuisance or anything, but could you add a validation check to my code and show it to me so I can get an idea as to how I can implement validation checks?
Explanations are good, but demonstrations are awesome.

[grinvader]

Complicated. You want to make sure every param/input value is the proper type and within boundaries, and if it's not, warn about it and stop execution or replace with proper defaults.
You also might want to warn in case there are too many or too few params. If you don't have defaults, too few must make your program stop gracefully, and so on.

Like...

Code: Select all

$ ./prog <integer> <filename1> <filename2>

Let's say prog reads 'filename1' and copies it 'integer' times into a file named 'filename2' (useless as it gets, but it's just an example)
Depending how you convert it from string to integer, you might have to check if integer is positive or not.
Filename1 has to exist to be read. No defaults make it that the program just exits saying "no input file, moron" if you don't give one.
Filename2 might exist, so you'll want added checks to overwrite the file if it does.
Oh, these 2 names must also be stored in buffers long enough to fit them. Each OS has his own limit for that.
If less than 3 params are passed on the CL, you want to either stop or ask for the missing ones (yeah right). If more than 3, depends your take on things.

That's just for those. Each input/param must be sanitized before you can trust it, else you open yourself to overflow exploits and stupid shit like that.

[ZH/Franky]

Well grin, thanks for your advice. I really do need to read up more on these things (also, I don't yet know how to do file input/output).

In any case, I've been messing around with for loops in vb lately.
I decided to rewrite that bin2dec program I posted earlier, in a much cleaner and efficient manner, and this is the result:

Code: Select all

Module Module1
    Private BinaryString As String
    Private digit As Double
    Sub Main()
        bin2dec()
    End Sub
    Private Sub bin2dec()
         Do
            BinaryString = ""
            digit = 0
            Console.WriteLine("Enter a binary number to convert to decimal: ")
            BinaryString = Console.ReadLine()
            For pos = BinaryString.Length To 1 Step -1
                If Mid(BinaryString, pos, 1) = "1" Then
                    digit = digit + (2 ^ (BinaryString.Length - pos))
                End If
            Next
            Console.WriteLine(digit)
         Loop
    End Sub
End Module

Man, the bin2dec program I initially posted was a complete mess.
I'm going to try do the same with my dec2bin program.
EDIT: Here's the new version of my dec2bin program:
(This one was a little harder to cleanup, but it's still cleaner than the last version I posted)

Code: Select all

Module Module1
    Private BinaryString As String
    Private digit As Double
    Private remain As Double
    Private binr As String
    Sub Main()
        Do
            dec2bin()
        Loop
    End Sub
    Private Sub dec2bin()
        Console.WriteLine("Enter decimal number to convert to binary: ")
        digit = Console.ReadLine()
        BinaryString = ""
        binr = ""
        Do Until digit = 0
            remain = (digit / 2) - Int(digit / 2)
            If remain > 0 Then
                binr = binr + "1"
            End If
            If remain = 0 Then
                binr = binr + "0"
            End If
            digit = (digit / 2) - remain
        Loop
        For pos = binr.Length To 1 Step -1
            BinaryString = BinaryString + Mid(binr, pos, 1)
        Next
        Console.WriteLine(BinaryString)
    End Sub
End Module

Screw while loops. For loops are much more powerful.

[funkyass]

VB. ugh.

while loops are good for when you have no idea the number of loops before entering the loop.

[ZH/Franky]

VB. ugh.

Well, it's the main programming language that I am supposed to use in my first year at college. Not like I have a choice, but then I also don't have any beef with VB.

What's wrong with VB? (If you don't mind that I ask)

[creaothceann]

It's a bad copy of Delphi.


EDIT: PM sent.

[ZH/Franky]

Delphi... you mean pascal?

Anyway, I merged both programs (bin2dec and dec2bin):

Code: Select all

Module Module1

    Private BinaryString As String
    Private digit As Double
    Private remain As Double
    Private binr As String
    Private choice As String

    Sub Main()
        Do
            Console.WriteLine("bin2dec or dec2bin?")
            choice = Console.ReadLine()
            If choice = "bin2dec" Then
                BinaryString = ""
                digit = 0
                Console.WriteLine("Enter a binary number to convert to decimal: ")
                BinaryString = Console.ReadLine()
                bin2dec()
                Console.WriteLine(digit)
            End If
            If choice = "dec2bin" Then
                Console.WriteLine("Enter decimal number to convert to binary: ")
                digit = Console.ReadLine()
                BinaryString = ""
                binr = ""
                dec2bin()
                Console.WriteLine(BinaryString)
            End If
        Loop
    End Sub

    Private Sub dec2bin()
        Do Until digit = 0
            remain = (digit / 2) - Int(digit / 2)
            If remain > 0 Then
                binr = binr + "1"
            End If
            If remain = 0 Then
                binr = binr + "0"
            End If
            digit = (digit / 2) - remain
        Loop
        For pos = binr.Length To 1 Step -1
            BinaryString = BinaryString + Mid(binr, pos, 1)
        Next
    End Sub

    Private Sub bin2dec()
        For pos = BinaryString.Length To 1 Step -1
            If Mid(BinaryString, pos, 1) = "1" Then
                digit = digit + (2 ^ (BinaryString.Length - pos))
            End If
        Next
    End Sub

End Module

All very efficient and easy to understand.
I've seperated the actual algorithms from the input/output commands, which makes it easy for me to implement new things now (such as the ability to convert negative integers to two's compliment binary, do binary addition, binary subtraction, binary multiplication, binary division, etc. Binary square roots! Oh yeah, you know that's cool.)


EDIT:
Not yet perfect, but I made an attempt at implementing negative decimal to two's compliment binary conversion:

Code: Select all

Module Module1

    Private BinaryString As String
    Private digit As Double
    Private remain As Double
    Private binr As String
    Private choice As String
    Private quit As Double
    Private negative As Double

    Sub Main()
        quit = 0
        Do Until quit = 1
            Console.WriteLine("bin2dec or dec2bin? to quit, enter q")
            choice = Console.ReadLine()
            If choice = "q" Then
                quit = 1
            End If
            If choice = "bin2dec" Then
                Console.WriteLine("Enter a binary number to convert to decimal: ")
                BinaryString = Console.ReadLine()
                digit = 0
                bin2dec()
                Console.WriteLine(digit)
            End If
            If choice = "dec2bin" Then
                Console.WriteLine("Enter decimal number to convert to binary: ")
                digit = Console.ReadLine()
                If digit < 0 Then
                    digit = digit * -1
                    negative = 1
                End If
                BinaryString = ""
                binr = ""
                dec2bin()
                If negative = 1 Then
                    binr = ""
                    flipbits()
                    binr = ""
                    bin2dec()
                    digit = digit + 1
                    dec2bin()
                    binr = ""
                    revbin()
                    negative = 0
                End If
                Console.WriteLine(BinaryString)
            End If
        Loop
    End Sub

    Private Sub dec2bin()
        Do Until digit = 0
            remain = (digit / 2) - Int(digit / 2)
            If remain > 0 Then
                binr = binr + "1"
            End If
            If remain = 0 Then
                binr = binr + "0"
            End If
            digit = (digit / 2) - remain
        Loop
        For pos = binr.Length To 1 Step -1
            BinaryString = BinaryString + Mid(binr, pos, 1)
        Next
    End Sub

    Private Sub bin2dec()
        For pos = BinaryString.Length To 1 Step -1
            If Mid(BinaryString, pos, 1) = "1" Then
                digit = digit + (2 ^ (BinaryString.Length - pos))
            End If
        Next
    End Sub

    Private Sub flipbits()
        binr = ""
        For pos = 1 To BinaryString.Length Step 1
            If Mid(BinaryString, pos, 1) = "1" Then
                binr = binr + "0"
            End If
            If Mid(BinaryString, pos, 1) = "0" Then
                binr = binr + "1"
            End If
        Next
        BinaryString = binr
    End Sub

    Private Sub revbin()
        For pos = BinaryString.Length To 1 Step -1
            binr = binr + Mid(BinaryString, pos, 1)
        Next
        BinaryString = binr
    End Sub

End Module

I'm not sure whether it's correct though. Am I right in saying that to find a negative decimal in binary you (for example, if you had 12, which in binary is 1100, and you wanted to find the binary value of -12) you would flip the zero's, in effect making 0011, or just 11, and then add 1, in effect creating 0100; that is, -12 = 0100?

As a side note, I rewrote that factorial finder I posted:

Code: Select all

Module Module1

    Private digit As Double

    Sub Main()
        Do
            Console.WriteLine("Enter number to find factorial of:")
            digit = Console.ReadLine()
            factorials()
            Console.WriteLine(digit)
        Loop
    End Sub

    Private Sub factorials()
        If digit > 0 Then
            For pos = digit To 2 Step -1
                digit = digit * (pos - 1)
            Next
        End If
        If digit < 0 Then
            For pos = digit To -2 Step 1
                digit = digit * (pos + 1)
            Next
            If digit > 0 Then
                digit = digit * (-1)
            End If
        End If
    End Sub

End Module

while loops are good for when you have no idea the number of loops before entering the loop.

Even then, I've found that "loop until" is often much better than "while condition... loop".
E.g.
while loop:

Code: Select all

while whatever = 1 do
   blah blah blah
   something = something + 1
   if something = somethingelse then
     whatever = 0
  end if
loop

"loop until" loop:

Code: Select all

loop until something = somethingelse
   blah blah blah
   something = something + 1
next

for loop:

Code: Select all

for pos = something to somethingelse step 1
   blah blah blah
loop

Maybe while loop's are cool for certain things, not sure; but I do know that recently, I've found myself writing much cleaner code by not using while loops.

[funkyass]

Look into the kinds of integers found in computer science. Here's a hint: are two, signed and unsigned. Once you get outside of whole numbers you are pretty much stuck using a separate set of functions for integers, and another set for real numbers for base conversions.

I even won't think of imaginary numbers.

As for loops, the major difference between structures is when you are checking the loop condition - before or after the block of code. Out of all programing languages, you are guaranteed while as part of its syntax(either by itself, or as a part of do-while variant).

[ZH/Franky]

Look into the kinds of integers found in computer science. Here's a hint: are two, signed and unsigned. Once you get outside of whole numbers you are pretty much stuck using a separate set of functions for integers, and another set for real numbers for base conversions.

Well, I think I remember reading that an unsigned integer is one that can be between 0 and 255 (that is, it's within an 8-bit address space), and that a signed integer is one that that can be between -255 and +255. That is, for example, 28 in an 8-bit address space would be 00011100, and -28 would be 11100100.
Am I correct?
So is "unsigned and signed" limited to 8-bit? Or does it have the be specified exactly how many bits you want to work with?
It all looks quite easy to understand; I just need to clear up some of the confusion.

[creaothceann]

The integers are the set of numbers consisting of the natural numbers including 0 (0, 1, 2, 3, ...) and their negatives (0, −1, −2, −3, ...). They are numbers that can be written without a fractional or decimal component, and fall within the set {... −2, −1, 0, 1, 2, ...}. For example, 65, 7, and −756 are integers; 1.6 and 1½ are not integers.

Integers can be signed or unsigned. "signed" means that half the range is in the negative domain.

Code: Select all

       byte =  8-bit integer: range of 2^8  =        256 values
       word = 16-bit integer: range of 2^16 =      65536 values
double word = 32-bit integer: range of 2^32 = 4294967296 values


unsigned  byte =           0..       255
unsigned  word =           0..     65535
unsigned dword =           0..4294967295

  signed  byte =        -128..127
  signed  word =      -32768..32767
  signed dword = -2147483648..2147483647

[Gil_Hamilton]

Look into the kinds of integers found in computer science. Here's a hint: are two, signed and unsigned. Once you get outside of whole numbers you are pretty much stuck using a separate set of functions for integers, and another set for real numbers for base conversions.

Well, I think I remember reading that an unsigned integer is one that can be between 0 and 255 (that is, it's within an 8-bit address space), and that a signed integer is one that that can be between -255 and +255. That is, for example, 28 in an 8-bit address space would be 00011100, and -28 would be 11100100.
Am I correct?

*shakes head*
Think of the range. An 8-bit sequence can represent 256 unique states.
-255 to +255 represents almost twice as many states. You need a 9th bit for -255 to +255.
9-bit values are highly uncommon, for various reasons.

An 8-bit signed integer represents -128 to +127.

So is "unsigned and signed" limited to 8-bit? Or does it have the be specified exactly how many bits you want to work with?

http://en.wikipedia.org/wiki/Integral_d ... type_names

While not limited to 1-byte values, the specifics depend on the language.

[creaothceann]

And the architecture (16-, 32- or 64-bit CPU).