Just wondering if I'm likely to encounter problems loading the multiple instances of the same module into my main application since each module's core needs a unique key, right? I figured I could somehow create a unique-random string for it's key (maybe based off the time?) when instantiated but this seems a little hacky.

My module is really a 'layout' which will be populated with content provided by XML when it's loaded into the shell so I want to re-use it multiple times.

This is a Flash CS4 app, not Flex.

I'm thinking of implementing the Pipes Utility into an Flash application (not Flex) I'm building for work. The application needs to be modular with each "page" being contained in it's own SWF. Each module will be unique to each other, one is a video player another an interactive map. Modules need to communicate with the shell and vice verca.

As each module is unique they generally each have a different API which the shell needs to communicate with. For example, the map module needs to respond to input passed down from the shell where as the video player only needs to alert the shell that it's finished playing or has been closed by the user.

Is this a good candidate for using the Pipes Utility? If a module is unloaded from the shell and a notification is attempted to be routed to it will an error occur or does it simply fail silently?

Any help/advice is appreciated 

I'm wondering what is the 'best practice' for handling value objects in an application, specifically where they should be stored.

In the past I have kept value objects inside a directory named vo under my model directory (model/vo/). This makes sense to me as VO's are used to transport data from the model to somewhere else in the application.

However I have a situation where I have one mediator which passes some 'view state' information as a VO to another mediator via a notification. Since this state data doesn't pass through the model tier should I create a separate vo directory under the view directory (view/vo/)? Or is it OK to simply continue using the currently structure of keeping all VO's in the model (model/vo/)?

I'm just wondering if anyone has attempted to create a Proxy for BulkLoader?

I was thinking of creating one but thought I'd ask here first. Basically the proxy would allow a series of items to be loaded as a group and then fire off a notification when finished. The notification would carry the loaded assets as it's body.

Has anything like this already been created? What's everyone using for bulk loading of assets?