Why is XML serialzation suggested over SOAP and binary serialzation?

Question

I'm trying to serialize a string with some escape characters (\0). But when I deserialise it throws an exception.

//XML serialization

    private static void M1()
        {
            string str = "\0AC";

           StringWriter sw = new StringWriter();
            XmlSerializer serializer = new XmlSerializer(typeof(String));
            serializer.Serialize(sw, str);
            System.Console.WriteLine("String encoded to XML = \n{0} \n", sw.ToString());
            StringReader sr = new StringReader(sw.ToString());
            String s2 = (String)serializer.Deserialize(sr);

            System.Console.WriteLine("String decoded from XML = \n {0}", s2);
        }

//Using Binary serialization

    void BinaryFormat1()
     {
         IFormatter bf = new BinaryFormatter();
         string str = "\0MyName";
         FileStream fs = new FileStream("SerilizedString.Data", FileMode.OpenOrCreate, FileAccess.ReadWrite, FileShare.ReadWrite);
         bf.Serialize(fs, str);
         fs.Close();
         fs = new FileStream("SerilizedString.Data", FileMode.Open, FileAccess.Read, FileShare.None);
         string data = (string)bf.Deserialize(fs);
         fs.Close();
         Console.WriteLine("data : " + data.ToString());
     }

Answer 1

I'd expect SOAP to have the same problem, to be honest. It's the issue of not being able to represent every Unicode string in "standard" XML.

That leaves binary serialization, which:

• Will only work with .NET (so is bad for interoperability)
• Isn't human readable (or easily translated into a human readable form if you run into problems)
• Is hard to work with in terms of versioning

Personally I prefer custom serialization options such as Thrift or Protocol Buffers, but as the author of a Protocol Buffers port to C#, I'm biased :)