> On Jun 2, 2016, at 8:55 AM, Daryle Walker <dary...@mac.com> wrote: > > [Warning: rambling] > > In Objective-C, you pretty much have to use a class for you model (in your > MVC Cocoa app). But in Swift, you have the option to use a struct/enum or a > non-NSObject class too. > > My model in mind is dumb data, so a struct seems appropriate. But your > various Cocoa subclass would need to reference the model, and using a struct > means it’s by value instead of reference, so coordinating changes would get > harder. > > I’m guessing that I could use something like view-models, and only those VMs > access the model, and always through referencing the containing object (like > a NSDocument subclass). But a reference type, even a non-NSObject one, still > seems easier. And I may want to use KVO or Core Data, which require NSObject > subclasses.
In practical terms, a model typically both has value and reference types. If you take a game example, the data that describes the capabilities of a unit is value data, but the unit itself has identity and is thus a reference type. Also “Dumb” vs “Smart” data doesn’t really translate well to value vs reference – you can have “smart” structs and “dumb” classes (a Box class is a typical example of a “dumb" class – its entire purpose is to grant identity to a value type) – its really a matter of if you need identity or not. I would highly recommend you experiment and do what feels for your case. -- David Duncan _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
