diff --git a/include/nana/charset.hpp b/include/nana/charset.hpp
index 41fa4d72..09f79bdf 100644
--- a/include/nana/charset.hpp
+++ b/include/nana/charset.hpp
@@ -46,7 +46,31 @@ namespace nana
 		class charset_encoding_interface;
 	}
 
-	/// An intelligent charset class for character code conversion.
+	/*!\class charset
+	\brief An intelligent charset class for character code conversion.
+	Example:
+	1. A UTF-8 string from the socket.
+
+	    int len = ::recv(sd, buf, buflen, 0);
+	    textbox.caption(nana::charset(std::string(buf, len), nana::unicode::utf8));
+
+	2. Send the string in text to the socket as UTF-8.
+
+	    std::string utf8str = nana::charset(textbox.caption()).to_bytes(nana::unicode::utf8);
+	    ::send(sd, utf8str.c_str(), utf8str.size(), 0);
+
+	3, Convert a string to the specified multi-byte character code.
+
+	    // Convert to a multibytes string through default system language.
+	    std::string mbstr = nana::charset(a_wstring);
+
+	    // If the default system language is English and convert
+	    // a Chinese unicode string to multibytes string through GB2312
+	    std::setlocale(LC_CTYPE, "zh_CN.GB2312");
+	               //set::setlocale(LC_CTYPE, ".936"); call it in Windows
+	    std::string mbstr = nana::charset(a_wstring_with_chinese);
+
+	*/
 	class charset
 	{
 	public:
@@ -74,27 +98,3 @@ namespace nana
 }//end namespace nana
 #endif
 
-/*!\class charset
-
-Example
-1. A UTF-8 string from the socket.
-
-		int len = ::recv(sd, buf, buflen, 0);
-		textbox.caption(nana::charset(std::string(buf, len), nana::unicode::utf8));
-
-2. Send the string in text to the socket as UTF-8.
-
-		std::string utf8str = nana::charset(textbox.caption()).to_bytes(nana::unicode::utf8);
-		::send(sd, utf8str.c_str(), utf8str.size(), 0);
-
-3, Convert a string to the specified multi-byte character code.
-
-			//Convert to a multibytes string through default system language.
-		std::string mbstr = nana::charset(a_wstring);
-			//If the default system language is English and convert
-			//a Chinese unicode string to multibytes string through GB2312
-		std::setlocale(LC_CTYPE, "zh_CN.GB2312"); 
-			//set::setlocale(LC_CTYPE, ".936"); call it in Windows
-		std::string mbstr = nana::charset(a_wstring_with_chinese);
-
-*/
\ No newline at end of file
diff --git a/source/charset.cpp b/source/charset.cpp
index 686f8182..95753662 100644
--- a/source/charset.cpp
+++ b/source/charset.cpp
@@ -578,37 +578,38 @@ namespace nana
 			std::string data_for_move_;
 		};
 #else
+        /// return the first code point and move the pointer to next character, springing to the end by errors
 		unsigned long utf8char(const unsigned char*& p, const unsigned char* end)
 		{
 			if(p != end)
 			{
-				if(*p < 0x80)
+				if(*p < 0x80)        // ASCII char   0-127 or 0-0x80
 				{
 					return *(p++);
 				}
 				unsigned ch = *p;
 				unsigned long code;
-				if(ch < 0xC0)
+				if(ch < 0xC0)       // error? - move to end. Posible ANSI or ISO code-page 
 				{
 					p = end;
 					return 0;
 				}
-				else if(ch < 0xE0 && (p + 1 <= end))
+				else if(ch < 0xE0 && (p + 1 <= end))      // two byte chararcter
 				{
 					code = ((ch & 0x1F) << 6) | (p[1] & 0x3F);
 					p += 2;
 				}
-				else if(ch < 0xF0 && (p + 2 <= end))
+				else if(ch < 0xF0 && (p + 2 <= end))     // 3 byte character
 				{
 					code = ((((ch & 0xF) << 6) | (p[1] & 0x3F)) << 6) | (p[2] & 0x3F);
 					p += 3;
 				}
-				else if(ch < 0x1F && (p + 3 <= end))
+				else if(ch < 0x1F && (p + 3 <= end))   // 4 byte character
 				{
 					code = ((((((ch & 0x7) << 6) | (p[1] & 0x3F)) << 6) | (p[2] & 0x3F)) << 6) | (p[3] & 0x3F);
 					p += 4;
 				}
-				else
+				else    //  error, go to end
 				{
 					p = end;
 					return 0;