#Original [package org.json] Douglas Crockford

The key changes are:

* Ensures order of elements

* Enables the normal HashMap methods to be available although most of them were overridden in earlier implementation like keys() and keySet()

* This in combination with JSONObjects extends from LinkedHashMap enables the whole library to be usable in a generic way where other libraries which uses model data in any combination of Map<String, Object> and List<Object> for example Map<String,List<Object>>.

* Attributes are marked with a '@' prepended to the key

* Content element is marked with a '#' prepended to the key

* Since the json keeps order of the original xml elements it can be deserialized back to same xml.

An xml of format

<root><person fname="samarjit" lname="samanta" >normal text</person></root>

//will serialize to

//In earlier version it used to serialize to

{"root":{"person":{"fname":"samarjit","lname":"samanta","content":"normal text"}}}

Since we have the attribute marked in the json, we can create back the exact xml

//This version if input json is

{"root":{"person":{"@fname":"samarjit","@lname":"samanta","#content":"normal text"}}}

Example of libraries that uses model as combination of Map<String,Object> and List<Object> and arbitrary java beans.

String templateExpression = "Hi ${ddd} hello ${ar[0]} your home is ${USER_HOME}";

System.out.println(ret);

```

Result

//This version produces

Hi jsss hello jhaldia your home is C:/Users/Samarjit

//Earlier version json:

Hi jsss hello [ your home is C:/Users/Samarjit

```java

JSONObject jobj1 = new JSONObject();

jobj1.put("ddd","jsss");