If you want your shell to be as lightwieght as possible, then you have to use a shared code module, or code up your own late-loading RSL scheme. Otherwise, you have to add the shared code to the main app which makes it fatter, but is much easier to code. In theory, the shared code module would include every class in common between your modules, although I can imagine scenarios where you'd have several shared code modules so there isn't one huge shared code module. The idea here is to only load a class once. If I were doing it, I'd write some sort of XML parser to read in link-reports and spit out the list of common classes. Maybe you can take the lead on making such a utility and sharing with others. -Alex
________________________________ From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of ben.clinkinbeard Sent: Wednesday, April 18, 2007 7:46 AM To: [email protected] Subject: [flexcoders] Shared code modules - how do you know what they need to contain? I am preparing to break up a monolithic app into modules, and have seen plenty of references to using shared code modules for things like managers. In my case, the shell application will be very basic and more or less just serve as a dashboard for navigating into various workflow engines. 1) Are shared code modules a virtual necessity? 2) How do you know what code needs to go into them? I have seen the examples on Alex Harui's site and while they're great to see how to structure the code, I still feel like I am lacking some of the "why" and "when" knowledge. Thanks, Ben
