Silverlight RPN Calculator Tutorial.

Well I beat my old nemesis and finished an RPN calculator. It was much easier than I remember, but then again, I tried it like four years ago. I've learned so much in the time since, I can hardly believe how silly my old code is.

Have any of you ever used a Reverse Polish Notation calculator? I did in high school. It was easily the best calculator ever (the HP=32SII). RPN is great, because you don't have to use parenthesis. The stack and the order in which you key the operators maintains order of operations.

Today, I'll walk you through creating one. Note this tutorial is technically for Silverlight, but except for the marked sections, it can be applied wholesale to WinForms or WPF. Most if this is simple stack logic, and that exists on all .NET platforms.

Read more after the jump.

Postfix expression parser.

A question on DreamInCode lead me to my old nemesis, the Postfix expression parser. I'd fallen in love with a Reverse Polish Notation calculator in high school (the HP-32SII, they sadly don't make them anymore). When I tried to program a RPNC to use on my PC years ago, I failed miserably. But I've learned a lot since then, and I figured I could at least try a simple postfix parser.

It wasn't that bad. You don't have to check for a lot of things a calculator does. Here's the body of the code. I made it static, since you really don't need to instantiate it. If I ever make it into a calculator, I'd probably make it non-static, and make the stack a class-level field.

class PostfixParser
{
    const string Placeholder = "[x]";

    private static Dictionary<string, Func<double, double, double>> op;

    private static List<string> operators = new List<string>() { "+", "-", "*", "/" };

    static PostfixParser()
    {
        op = new Dictionary<string, Func<double, double, double>>();
        op.Add("+", (x, y) => x + y);
        op.Add("-", (x, y) => x - y);
        op.Add("*", (x, y) => x * y);
        op.Add("/", (x, y) => x / y);
        op.Add("^", (x, y) => Math.Pow(x, y));
    }

    public static double ParseExp(string exp, double? val)
    {
        string[] tokens = exp.Split(new string[] { " " } , StringSplitOptions.RemoveEmptyEntries);
        if (tokens.Contains(Placeholder) && val == null)
            throw new InvalidOperationException("Placeholder token detected in expression, but parameter \"val\" was null.");
        Stack<double> stack = new Stack<double>();
        foreach (string token in tokens)
        {
            if (operators.Contains(token))
            {
                if (stack.Count < 2)
                    throw new InvalidOperationException("Operator imbalance.");
                double y = stack.Pop();
                double x = stack.Pop();
                stack.Push(op[token](x, y));
            }
            else
            {
                double n;
                if (token == Placeholder)
                    n = val.Value;
                else if (!double.TryParse(token, out n))
                    throw new InvalidOperationException("Invalid token detected: " + token);
                stack.Push(n);
            }
        }

        if (stack.Count > 1)
            throw new InvalidOperationException("More than one result on the stack.");
        if (stack.Count < 1)
            throw new InvalidOperationException("No results on the stack.");
        return stack.Pop();
    }

    public static double ParseExp(string exp)
    {
        return ParseExp(exp, null);
    }
}

This can be used like this:

string fToCExp = "5 9 / [x] 32 - *";
string cToFExp = "9 5 / [x] * 32 + ";
Console.WriteLine("100C converted to F:");
Console.WriteLine(PostfixParser.ParseExp(cToFExp, 100));
Console.WriteLine("32F converted to C:");
Console.WriteLine(PostfixParser.ParseExp(fToCExp, 32));
Console.ReadKey();

To give an output of:

100C converted to F:
212
32F converted to C:
0

You could add more operators if you wished, like a 1/x or a SQRT, but since most other operators are unary rather than binary, you'd have to change your popping logic.

Note: I did it with a placeholder value in mind. I used "[x]" for whatever reason, but it could be anything that won't parse into a double by itself or match one of the operators. You could remove this altogether, if you cared to.