Hi guys! I am writing an MMO (crazy, I know), and I need to figure out the best way to create actions for items. I way I think would be the best would be to create an XML file for each object, and have a section of that file that contains some code to be loaded into the client when the object is accessed, sort of a temporary plugin. The code would just be a type of script that would call basic functions on the server to avoid hard-coding anything.

What do you guys think of this approach? Would it be too susceptible to hacking if it runs from the client? Would it be easier to have an "object server" of some type, so the client just passes the name of the item & action, and the server does the rest? Or am I looking at this completely wrong?

Thanks for any input you have!
SphereCat1

Could you supply an example of an action for an item? Usually Items don't perform actions. Are you talking about a "USE" functionality where a player causes an action by Using an Item?
If that is the case, the USE functionality is usually built into the class that is doing the using. static attributes particular to the item belong in the class of the item.


As a general rule, you don't want to store any behavior or game state related data on the client. The ability to modify and change this data is too difficult to contain.

hope that helps.
Aet

Yeah, the use functionality is kind of what I'm going for. So that each item doesn't have to have it's effect hard-coded into the server.

But I'm also trying to get a bigger scope for the functionality, so that if there was a space station, for instance, it could have a self-destruct function built into it, as well as move, spin, etc. So I guess object would be a better word for it than item.

Thanks for the quick reply!

You would normally have an Item class, and then a class for each item which inherits the Item class.

There will be a method called Use in the Item class that gets overloaded in the child class, which has code which is executed when that item is used.

In a single-player game, the client will use the item, the item will be destroyed and the inventory will be updated, and the code inside the Use() method will go off.

In an MMO, 1 of 2 things will happen:

1) The client will use the item, destroy it, update inventory, and then send a message to let the server know that the item has been used OR

2) A REQUEST will be sent from the client to the server to let the server know the client would LIKE to use the item, the server will check to see that this is a valid move, then send back to the client whether or not the client can actually perform this request.

Depending on the scale of the game, it is better to go with Option 2. Most other actions fall under Option 2 as well.

So then I would just store the child classes in separate files that can be swapped in and out when necessary? That's a good idea. I could create a wrapper file for an item that contains a Class file, and also all the 3D models and textures needed to create an item. Then it's just a matter of passing the models to the client. Sounds like a plan!