Hi
I'm no game coding expert, but I've seen and had to revise many home grown -spaghetti code game engines.

For a simple(non-database) game, where you need to track and increase the score, notify the user(you lose, you win etc) that requires basic keyboard interaction, would puremvc be overkill?

If not could someone suggest 'best practice' for architecting?

What logic if any would be in the model, or would the model simply be maintaining application state?
How much logic to put in controllers vs views, etc?

Any feedback appreciated!

Hi
I'm no game coding export, but I've seen and had to revise many home grown -spaghetti code game engines.

For a simple(non-database) game, where you need to track and increase the score, notify the user(you lose, you win etc) that requires basic keyboard interaction, would puremvc be overkill?

If not could someone suggest 'best practice' for architecting?

What logic if any would be in the model, or would the model simply be maintaining application state?
How much logic to put in vicontrollers vs views, etc?

Any feedback appreciated!

Not sure I agree with Apple's approach but my guess is that it has a lot to do with the template-like nature of most iphone applications. If the majority of your view handling logic sits in the viewcontroller, then it really defeats one of the advantages of MVC, keeping the view components fully modular so they can be swapped out and interchanged.

But perhaps someone with more framework savvy can chime in and tell me why I'm wrong.
Also, I guess maybe its time to have a peak at the ObjC port of PureMVC to see if I can apply it in my projects.

Having gone thru most of Apple's Objective-C documentation on 'View Controllers', I am somewhat confused by how this term is used.
Objective C descends directly from Smalltalk so am I correct in assuming that a viewcontroller as used in Objective C is a more 'pure' implementation of what a Controller object should be in any MVC based application?

Initially, in coming from an AS3 and PureMVC framework, I though of ObjC view Controller's as being like composite views that also encompassed some of the responsibility of mediators in a PureMVC based application (handling of initialization, hiding and showing of views)

Has anyone else though about this and has working with other languages with MVC based frameworks influenced how you distribute your logic and responsibilities in Actionscript based projects?

Work in progress.

Inspired by Bitmap genius Quasimondo(Mario Klingemann) at FITC this year.
Some adaptive subsampling niceness...or 'playing with bitmaps'
PureMVC mutlicore
http://bitstream.ca/imageSubSample/

Hi
I use and love PureMVC. Its been the cornerstone for over half a dozen large commercial websites I've worked on the past year.
But as a serious developer I'm always on the look out for better ways to achieve any given task. I was wondering if anyone has successfully integrated Robert Penner's new AS3 signals with a PureMVC project?

I think for my next project I will use a mediator.

I don't have time to refactor because of a deadline.

Option A wouldn't work for me because my load is broken down into 2 loads, and BOTH need notifications from swf Address immediately after the application inititializes, browser loads.

I've resorted to a polling method in my SWFAddress proxy, which checks for a 'loadCompleted' flag to be true. Only then does it send the notification.

This flag in the SWFAddress proxy is set via a command mapped to a LOADCOMPLETE notification sent from my main DataProxy, which handles all loading, parsing of data.

Hi
In my current architecture I am using a proxy to handle all swfadddress requests.

It works fine except for browser url changes, which trigger  a SWFAddressEvent.CHANGE event, which in turn causes a notification to be sent.

Its a typical chicken & egg problem, the notification is sent BEFORE the view mediators have been registered.

If I kludge it by adding a half second delay it works.

What I'd like to do is instead start a polling emthod in the SWFAddress proxy, which checks/listens for all assets to be loaded BEFORE sending the notification.

What is the best way to handle proxy to proxy communication?

Hi
I have a completed project entirely contained in 1 swf.
I'd like to break it down partially, removing a few of the view components and loading them at runtime.

These view components aren't referenced directly by any mediator, but rather are parented and instantiated by other view components.

What is the best (practices) strategy to do this?

Besides making my application more modular and easier to work on with other developers, one of my goals is also to reduce the size of my 'main' swf, so this means removing any references to the view component class I need to load.

Should I then create a view  interface for these loadable components, which the parent components reference and instantiate??

Any suggestions welcome...

I can only share my preferences for loadup strategy.
One of the things that really irks me is long preload times for flash-flex sites...I don't care how interesting, life changing,  or potentially award winning the content, if I have to sit and wait for more than around 10-15 seconds....well..life is simply too short, you've lost me as a viewer!
Maybe I'm too impatient, but I think most casual surfers feel the same. It's also given flash (and especially Flex) a bad reputation as a seemless user experience. As a developer I also feel that part of my job is to 'educate' the client, letting them know that a long preload time (some clients expect this) isn't necessary.

So, having said that, I try and target the size of my 'main' swf to be no larger than 2-300K...and preferably under 200k! In an ideal world, IMHO there would be NO PRELOADERS! (and no 'skip intro' buttons;)

To load this I also use a 'shell' swf, but it is simply that, a shell, absolutely no logic other than to display a preload bar and load my Main swf. Lee brimelow outlines this simple scenario and gives reasons for this as a best practice over on gotoandlearn.
If I need to load further assets once Main is loaded using BulkLoader or some other utility, I do so, but usually in the background so it is transparent to the user. There should be at least some functionality in the site so the user doesn't feel like they need to wait to do something.

Structuring a load process properly takes more effort and time, which is probably why there are so many viewers staring at preloaders hoping their wait is worth it!
Just my 1.5 cents;)

Can someone tell me where to get the non debug versions of PureMVC?
The link at the bottom of this page
http://trac.puremvc.org/PureMVC_AS3_MultiCore/wiki/Downloads
I believe is for the debug version (the swc is 12k)

Thanks Cliff...will have a look!

I know there is a Port to Objective C forum, but my question whether anyone has taken an application they've developed in as3, and then converted to ObjectiveC?

I'm talking about re architecting from scratch at the code level, not using CS5's ability to export to native iPhone code.

Can you share examples?
What were the pitfalls, gotchas?

With the iPad arrival, this is going to be a growing area of interest for many flash developers I'm sure.

A Twitter search appliance with some added features

In its first incarnation, mainly 3d widow dressing on top of the Twitter advanced search api.
But slowly adding new features to make it much more.

http://www.tweetmasher.com

puremvc multicore and fp10 3d api.

Hi
Is a URLProxy necessary(or best practice) for simply loading a URL in a browser window?
OR simple keep the logic in the view component?