My guess (just a wild one for not being a thinker of touch interface design), would be to solve the solution by first detecting what kind of flash app it is and how much area it takes up, then anytime the finger scrolls over that particular area, there would be a cursor (the user could turn this off, but this is what would happen by default). As the cursor is led by your one finger which never lets up, meanwhile 1,2, 3 or 4 taps with another finger along with other more complex combinations eventually could perform various flash actions. Contrary to Dilger's primary thesis, I believe it's pretty straight forward that Apple is not as limited in thinking to make it happen if they want to, and that even the general audience is ready to take another small step in the gestures department.
Why should Apple think hard and work hard in order to solve a really hard problem they have no particular interest in solving?
Maybe there is an elegant solution out there to make all or nearly all existing Flash apps work, but it won’t be easy and I don’t see Apple investing much in technology they clearly think is dying. They normally never do.
My point was that now that Flash does support multitouch and gestures, it's possible to add those features to existing content (yes, this will happen). Naturally it won't apply to all projects but it's far from the sinking ship that the author suggests and rabid HTML5 supporters are praying for.