if the plugin architecture is going away (and i’m not sure it is given that flash sits alongside java (not so impressive) and unity 3d (which seems like it could be the future of game demo software as well as web-based 3d games - see this for example)) then adobe will find another way (recent proof of concepts being the javascript engine to run flash and exporting to canvas).

The problem is that most of the arguments against flash are based on the fact that flash isn’t open and the web is, but then that makes it okay for apple to block cross compilation from flash to cocoa which is a hell of a leap. Adobe’s argument is that a ‘develop once, publish on all’ model are a better, more open way of development than bespoke development for each platform which is expensive, time consuming and ultimately out of reach for most (small) developers. And because Flash 10.1 is hardware accelerated on all open gl es 2.0 devices, which includes android and would include iphone, the battery and performance issues would be down to the hardware for the most part (fp10.1 hardware acceleration article).

The web is an excellent way of delivering content. That might be information (as with most websites), applications and games, and interactive experiences. Flash is good for delivering the latter 2 but is unfortunately used for all 3 sometimes (as a flash developer i make sure that when a clearly html site is briefed in to me, that i let the account manager know that flash is a mistake for all of the reasons that zeldman(?) gives) so the comparison has to be between flash and the appstore. Of those 2, flash is the most searchable (based on Google’s swf-indexing forays and the fact that all web-based flash is delivered on an html page which is indexable) and portable (in that the flash is the same no matter which browser or operating system you choose to view it on) of the currently viable technologies (this showing that html5 isn’t quite the flash killer that it’s cracked up to be… yet).