It's worth noting that Adobe is very reluctant to deprecate API's. They've historically preferred (with AVM1/AVM2) bundling new VM's into the player over deprecating features. And any deprecation must be version-specific, and tends to be more for API consistency than anything else.
What really disappoints me is Adobe's reluctance to support really neat Labs projects like Alchemy. Want to run C (or any language which compiles to C, like haskell!) in the browser? It's been possible since Flash 9, way before NaCl. Official support is coming soon but it's a bit too bad it took so long.
The (deprecated) via-c backend of GHC compiled to something that looked like C, but depended on a horrible Perl script to post-process the object output.
It does have an LLVM backend these days, though, maybe that can work?
I can't tell if you're saying you know or not, but Alchemy is becoming an official product:
http://blogs.adobe.com/flashplayer/2011/09/updates-from-the-...
WHAT? Are you serious? Please educate us as to how that is done.
An unhandled exception due to referencing a null object that caused the Flash app to crash? Developer's fault, not Adobe's.
A crash that brings down the entire Flash runtime (and possibly the browser hosting it)? Adobe(or the browser vendor)'s fault, not the developers.
Even if the developer does something horribly stupid in ActionScript, the runtime should never die due to developer error.
I honestly can't remember ever seeing a Flash crash actually bring down the the browser. Maybe I've been very lucky.
I see that I'm getting down voted for calling 'ED' out on his sweeping statement where he did not define what he meant by a "crash". Note that for many users a "crash" is when what ever App/game/web-page they are running stops with some error and they can't continue. My basic point is that many (most??) of these, in my experience, are down to developer error.
Many developers seem to have an irrational "hate-fest" going against Flash. For sure there's a lot wrong with it but, for me, all to often "its not the tool but the craftsman at fault".
E.g. - the developer asks for an runtime created asset that doesn't exist anymore or there's a condition under which a block of code runs forever (actually in this case the flash runtime stops after 15 seconds of activity and allows the user to kill it but it's still a "crash" insofar as the user is concerned.
I suspect Flash's (lack of) stability was one of the reasons for considering this design decision in Chrome.
Still, if you look at what they've done in the past couple of years you'd see a lot in trying to revert that situation. Forcing a maximum tick rate of 60hz (was 120), dropping memory use and performance (tick rate goes to 2hz when hidden on a separate tab, or out of the screen); coming up with new browser APIs so the browse can control the plugin performance; and on and on.
They do deprecate a lot over the years. However, is "eliminated" something people really want? The platform prides itself in the fact that SWFs that worked for Flash 2 still work today, despite the changes in the language and virtual machine. So unless it's a change of security policy, nothing breaks or is "eliminated"; as a developer, I find this to be a good thing as I don't have to have a client coming to me after 6 months complaining about something not working anymore.
Flash is a platform that allows people to do a lot. Including shitty binaries. And there's not much defense against it.
And personally I still have my browsers crash more because of shitty HTML/JS than Flash. YMMV.