Thanks Cliff. Thats a relief.

It would be helpful to see an example where this practice is used for larger applications. It would also be nice to see an example that interacts with some backend like Java, Ruby or PHP. I've been reading some things that is used with cairngorm (yuk), but they always have to define the models twice. Once on the server side, and another time on the client-side in flex. I'm having a hard time coming up with the distinction when its necessary to have server-side processing and client side since actionscript can do so much. It cannot be just for data storage.

Another thing is, as the application grows, it would be nice to have various flex application modules in individual swf's that are loaded as necessary. For example the first being just the foundation of users, groups, and security that loads additional resources as necessary.

Peter

Hello guys,

I've been looking into using PureMVC for a new project. It's clean and I like the concepts of Commands, Mediators and Proxies. But, what led to the design consideration to add the entire system's states as string constants in the application facade? In a system with many modules that are loosely coupled, for every module its commands and states will have to be listed in the facade. This list will become huge as my application grows. Wouldn't the best approach be to write each module in its own independent package so I could easily move it to other applications? That is one of the benefits of MVC. Even from a development perspective when working on teams, multiple people can be editing the same ApplicationFacade.as file to add a completely unrelated module to the system. I think it would be better to keep those sorts of things with the controller of those modules. Am I missing the point on something here?

Peter