Such Type of error generate due to inconsistency between the character encoding specified in the XML declaration and the actual character encoding used to serialize the XML document. In XML all characters come from the Universal Character Set (UCS), which always associates a numerical code point with each character. Many algorithms exist for converting code points into a sequence of bytes. A specific character encoding must be used to serialize an XML document. For example, if you type an XML document into notepad and save it, you can choose from one of several supported character encodings including ISO-8859-1, UTF-8, or UTF-16.
According to the XML 1.0 specification, all processors are required to automatically detect and support the UTF-8 and UTF-16 encodings.You don't need an XML declaration if you use one of these two encodings when serializing your documents,

If you use different encoding other than UTF-8/UTF-16, then you must use an XML declaration to specify the actual encoding used. Without knowing what encoding was actually used how can the processor read the encoding information?
It's easy for processors to auto-detect between UTF-8/UTF-16 with or without an XML declaration by looking for a byte order mark (BOM) required in UTF-16 documents. For all other encodings, you know that the first five characters must be "<?xml". Since a given processor will only support a finite set of encodings, a brute-force algorithm can be used that simply looks at the first few bytes to determine the family of the character encoding used (there are five possible encoding families including UTF-16 big endian, UTF-16 little endian, UCS-4 or other 32-bit encoding, EBCDIC, and everything else). Once the processor detects the encoding family, it can read the rest of the XML declaration (since only a restricted set of characters can be used in the XML declaration), then it can switch to using the specified character encoding within the detected family. If, at this point, the XML declaration tells the processor to switch to an encoding from a completely different family, that error occurs.
So any time you save a document using a specific encoding, then indicate a different encoding in the XML declaration, you will get this error because the encodings must agree.
 

Windows 2000 notepad files saved as Unicode use "UTF-16" encoding format.

If you add an encoding attribute to XML files saved as Unicode, windows encoding values will generate an error.

> The following encoding will NOT generate an error in IE 5+, Firefox, or Opera, but in Netscape 6.2 it will generate an error.