Some colleagues and I are drifting (I mean "striding resolutely") into the tail-end of a project involving considerable amounts of Flex code. This has led us to look back and try to map the best approach for upcoming Flex-centred projects (aka "Impending Doom"). Part of that discussion has involved punting around the need for and benefits of using an MVC framework with Flex. We've talked about scale and complexity as a driver, as well as client preferences and timelines... no definitive answer has been found, as of yet.
In the middle of all this, I came across this article, "MVC As Anti-Pattern" by Brian Lesser of Ryerson (a fine learning establishment that has escalators, for crying out loud!). As University Types are wont to do, he mentions Smalltalk a lot. If you can overlook that fact, however, you'll find that Brian's points are quite well put (even if you disagree).
(In honour of University Types everywhere, I will now endeavour to use as much Latin as I can in a single paragraph.)
My inadequate summary: the author plants the seeds of doubt as to the usefulness of formal MVC in Flex RIAs given that the Flex framework itself provides in situ constructs that cover most of what a controller does (ex generis: data binding, validation, program flow control). In the end, Flex lends itself better to "MV" -- id est: split out the Model and View, with the Controller implied but not really a concrete entity.
If you to are pondering the benefits of using MVC in designing your next Flex app, then please read this article and click through to the "References" links he includes. I welcome any opinions you'd like to voice on the matter, since I'm still in the middle of forming an opinion on the issue.
In the middle of all this, I came across this article, "MVC As Anti-Pattern" by Brian Lesser of Ryerson (a fine learning establishment that has escalators, for crying out loud!). As University Types are wont to do, he mentions Smalltalk a lot. If you can overlook that fact, however, you'll find that Brian's points are quite well put (even if you disagree).
(In honour of University Types everywhere, I will now endeavour to use as much Latin as I can in a single paragraph.)
My inadequate summary: the author plants the seeds of doubt as to the usefulness of formal MVC in Flex RIAs given that the Flex framework itself provides in situ constructs that cover most of what a controller does (ex generis: data binding, validation, program flow control). In the end, Flex lends itself better to "MV" -- id est: split out the Model and View, with the Controller implied but not really a concrete entity.
If you to are pondering the benefits of using MVC in designing your next Flex app, then please read this article and click through to the "References" links he includes. I welcome any opinions you'd like to voice on the matter, since I'm still in the middle of forming an opinion on the issue.