Title: moving from single core pureMVC to Multicore pureMVC Post by: amerineni on February 17, 2010, 09:26:58 Hi
we are planning to move to multicore PureMVC and divide each view into modules. I have the following questions 1) We are planning to have only one module loaded at a time. In this case, how can I approach In order to pass data from the existing module to the module being loaded(next module)? 2) Can we talk to the main application using the notifications or do we need to use the pipes utility? My understanding is that pipes is mainly used for communication between modules Which are available at the same time and for communication with the main shell? 3) We are currently not using proxy for storing data, but using for handling the result and fault handlers and having methods which can be called from mediators. Do you think it’s a good way If I use my shell proxy(Main applications proxy ) like a model locator in cairngorm with different attributes? 4) Any other best practices or suggestions using multicore pureMVC Your suggestions will be really helpful. Thanks in advance. Vin a Title: Re: moving from single core pureMVC to Multicore pureMVC Post by: puremvc on February 17, 2010, 10:38:52 1) We are planning to have only one module loaded at a time. In this case, how can I approach In order to pass data from the existing module to the module being loaded(next module)?I would suggest having a long-lived model module that each successive loaded module communicates with. 2) Can we talk to the main application using the notifications or do we need to use the pipes utility? My understanding is that pipes is mainly used for communication between modules Use Pipes and Messages if possible. Or interfaces and events. Otherwise you have the problem of maintaining your notification namespace between cores. 3) We are currently not using proxy for storing data, but using for handling the result and fault handlers and having methods which can be called from mediators. Do you think it’s a good wayIf you need to persist data between module loadings, you probably do want to store the results in the Proxy or you'll end up making the calls again, right? If I use my shell proxy(Main applications proxy ) like a model locator in cairngorm with different attributes? Then you use an interface to define the shell and all its methods (and properties can be governed by defining get and set methods on the interface) and all modules need a reference to the shell. It's not the best arrangement (See the answer to #1), but it works. See the Modularity demo for interface/event communications. 4) Any other best practices or suggestions using multicore pureMVC
-=Cliff> |