Modular programming tips?

I'm working on a system for interacting with a game world that needs to be able to be easily expanded upon for new features, yet wouldn't require the aquisition of new objects as new features come out. Essentially, I need to develop a core system that can have modules added to it later on without overwriting anything from the original.

right now I have no idea how to go about making a modular system, so my current plans are to have the objects save all it's data, and when a new module is introduced, it can regurgitate it's saved data to the new module and the new one will replace the old one.

It would be nice if I could just create a new script for the new module to be added, and somehow tie it's workings to the old script. Unfortunately, I'm not sure how I'd go about making the old script be able to take commands from a new module if it does something as drastic as, say, introduce a completely new play style to the game world.






*side note* One of the current big issues with the current designs is that all the important data is stored in each individual's item in script. Since any kind of "how to make an off-world storage system" makes no sense to me, off world storage is not an option at this point in time. It'd be nice if there was a simple windows program that you could run and it'd set up a data storage system for SL automatically and already have the protocols for receiving and sending the data in world >_>.

Hmm. That's an interesting architectural question. To some degree, it is going to depend on what this system is designed to do. If the answer is, "Well, anything," then it is not likely to have an easy answer and you might need just a general-purpose product update system like /54/b2/308491/1.html (though you'll still have to work out how to do that data save/load if it's what you need).

If, on the other hand, it is narrow enough to start talking about what different modules will have in common, how they might need to communicate with each other, etc., then we might be able to start coming up with some kind of plug-in type solution with truly modular scripts and a central framework of some sort.



That's a tempting problem to work on. Hmm. I might work on some sort of general-purpose owner+storageKey storage system similar to the Open SL Product Update Server program I linked to above; a web application in Perl or PHP and a template for building a LSL script that uses it. Hmm.

Senuka, take a look at w-hat's (They are the one's that host lslint) site for a free database setup for SL. You don't have to use that site(I've not heard of any complaints), but it is a good place to start. The page is pretty easy to understand:

http://w-hat.com/httpdb