Welcome to the demonstration of BigFloat for Tcl !
If you have problems with this demonstration, first see the file README.txt in the package.

The principle of the RPN calculator is that you have a stack of numbers ;
you can push in the stack either integers or floating-point numbers, with any number of digits.
Suppose you have pushed 1.00 and 2.00 (you have to type each number in the entry field at the middle of the window)
Then the stack looks like :
4 : Empty
3 : Empty
2 : 1.0
1 : 2.0

By clicking on the button named "add", the program takes the two numbers off the stack,
and pushes instead of them their sum.
The stack then looks like :
4 : Empty
3 : Empty
2 : Empty
1 : 3.0

Now push 5.00 into the stack, and click on the button named "sub".
The stack now looks like :
4 : Empty
3 : Empty
2 : Empty
1 : 2.0

The result is the difference between the last number and the last but one on the stack.
5 - 3 = 2

The "append zeros" label is followed by a field in which you may enter an integer.
When you want to type 1.00000000 to get much precision, it is better to type '1.' in the left field,
and to type '8' in the right field, and then to push the number on the stack.
The result is the first floating-point number followed by '8' zeros : '1.00000000'.
Some commands need an integer parameter. An integer is a number formed with digits and (optionally)
the minus sign.
Example : 12  20000  -15 are integers
Floating-point number are formed using digits, eventually a minus sign, and formally a dot or
an exponent character ('e').
Examples of floating-point numbers : 1.23e-13 2. 0.10000000 23e54

Floating-point numbers are represented with an arbitrary precision, but the last,
and eventually the last but one digit(s) are truncated in the result as
they specify an uncertainty on the number. When you mean 1.3, you may have to append zeros
at the end, which can be done with the second field.