diff --git a/js/security/Witness.js b/js/security/Witness.js
index 3b88f06..1a8c5e0 100644
--- a/js/security/Witness.js
+++ b/js/security/Witness.js
@@ -13,17 +13,119 @@
 	this.path = new MerklePath();  // MerklePath
 };
 
+function parseOID(bytes, start, end) {
+    var s, n = 0, bits = 0;
+    for (var i = start; i < end; ++i) {
+        var v = bytes[i];
+        n = (n << 7) | (v & 0x7F);
+        bits += 7;
+        if (!(v & 0x80)) { // finished
+            if (s == undefined)
+                s = parseInt(n / 40) + "." + (n % 40);
+            else
+                s += "." + ((bits >= 31) ? "bigint" : n);
+            n = bits = 0;
+        }
+        s += String.fromCharCode();
+    }
+    return s;
+}
+
+function parseInteger(bytes, start, end) {
+    var n = 0;
+    for (var i = start; i < end; ++i)
+        n = (n << 8) | bytes[i];
+    return n;
+}
+
 Witness.prototype.decode = function(/* Uint8Array */ witness) {
+	/* The asn1.js decoder has some bug and 
+	 * cannot decode certain kind of witness.
+	 * So we use an alternative (and dirty) hack
+	 * to read witness from byte streams
+	 *      ------Wentao
+	 */
+	/*
 	var wit = DataUtils.toHex(witness).toLowerCase();
-	var der = Hex.decode(wit);
-	var asn1 = ASN1.decode(der);
+	try {
+		var der = Hex.decode(wit);
+		var asn1 = ASN1.decode(der);
+	}
+	catch (e) {
+		console.log(e);
+		console.log(wit);
+	}
 	//console.log(asn1.toPrettyString());
 	
 	this.oid = asn1.sub[0].sub[0].content();  // OID
+	//console.log(this.oid);
 	this.path.index = asn1.sub[1].sub[0].sub[0].content();  // index
+	//console.log(this.path.index);
 	for (i = 0; i < asn1.sub[1].sub[0].sub[1].sub.length; i++) {
 		pos = asn1.sub[1].sub[0].sub[1].sub[i].stream.pos;
 		str = wit.substring(2 * pos + 4, 2 * pos + 68);
 		this.path.digestList.push(str);  // digest hex string
+		//console.log(str);
+	}
+	*/
+	
+	// FIXME: Need to be fixed to support arbitrary ASN1 encoding,
+	// But do we really nned that????  -------Wentao
+	
+	// The structure of Witness is fixed as follows:
+	// SEQUENCE  (2 elem)
+	//   SEQUENCE  (1 elem)
+	//     OBJECT IDENTIFIER  1.2.840.113550.11.1.2.2
+	//   OCTET STRING  (1 elem)
+	//     SEQUENCE  (2 elem)
+	//       INTEGER  index
+	//       SEQUENCE  (n elem)
+	//         OCTET STRING(32 byte) 345FB4B5E9A1D2FF450ECA87EB87601683027A1A...
+	//         OCTET STRING(32 byte) DBCEE5B7A6C2B851B029324197DDBD9A655723DC...
+	//         OCTET STRING(32 byte) 4C79B2D256E4CD657A27F01DCB51AC3C56A24E71...
+	//         OCTET STRING(32 byte) 7F7FB169604A87EAC94378F0BDB4FC5D5899AB88...
+	//         ......
+	// Hence we can follow this structure to extract witness fields at fixed level
+	// Tag numbers for ASN1:
+	//    SEQUENCE            0x10
+	//    OCT STRING          0x04
+	//    INTEGER             0x02
+	//    OBJECT IDENTIFIER   0x06
+	var i = 0;
+	var step = 0;  // count of sequence tag
+	while (i < witness.length) {
+		var len = 0;
+		
+		if (witness[i] == 0x30) {
+			// Sequence (constructed)
+			// There is no primitive sequence in Witness
+			if ((witness[i + 1] & 0x80) != 0) {
+				len = witness[i+1] & 0x7F;
+			}
+			step++;
+		} else if (witness[i] == 0x06) {
+			// Decode OID
+			len = witness[i+1];  // XXX: OID will not be longer than 127 bytes
+			this.oid = parseOID(witness, i + 2, i + 2 + len);
+			//console.log(this.oid);
+		} else if (witness[i] == 0x02) {
+			// Decode node index
+			len = witness[i+1];  // XXX: index will not be longer than 127 bytes
+			this.path.index = parseInteger(witness, i + 2, i + 2 + len);
+			//console.log(this.path.index);
+		} else if (witness[i] == 0x04) {
+			if ((witness[i + 1] & 0x80) != 0) {
+				len = witness[i+1] & 0x7F;
+			} else {
+				len = witness[i+1];
+			}
+			if (step == 4) {
+				// Start to decode digest hex string
+				str = DataUtils.toHex(witness.subarray(i + 2, i + 2 + len));
+				this.path.digestList.push(str);  // digest hex string
+				//console.log(str);
+			}
+		}
+		i = i + 2 + len;
 	}
 };