Creating Ordered Complex Datatypes

From AIBenchWiki

It is now possible to define a relative order inside COMPLEX Datatypes.

This allows for the correct ordering of subitems inside Datatypes. When serialization/marshalling operations occur, it is usual for AIBench to lose the original positioning of the elements inside the JTree. This extends the COMPLEX Datatypes with a persistent relative ordering for the subitems of a given Datatype.

To do it one must do the following:

• The @Clipboard annotation (inside COMPLEX Datatypes) now has a method called order(), by defining it, you will be defining the relative order of that element inside the Datatype.

@Clipboard(name="Name",order=1){
public String getName(){
   return this.name;
}

---

• The order is interpreted as a magnitude order, so you can define several elements with gaps, just as follows:

@Clipboard(name="x",order=1)
...
@Clipboard(name="y",order=10)
...
@Clipboard(name="z",order=100)

---

• The user must not repeat orders within Datatypes. AIBench will throw an IllegalArgumentException if that happens:

@Clipboard(name="x",order=1)
...
@Clipboard(name="y",order=1)
...
@Clipboard(name="z",order=10)

---

• The user is responsible to guarantee that all the @Clipboard annotated methods inside a COMPLEX Datatype are correctly ordered.

Meaning that, if one defines an order for a method inside a Datatype, but doesn't define the order for the other methods inside that same Datatype, the behavior will probably be incoherent.

---

• If one does not define any order for all the @Clipboard annotated methods inside a COMPLEX Datatype, the behavior will be the traditional one, meaning that the ordering is fully OPTIONAL.