So the new EITF rules will basically take the subscription-based model and apply it only to the value of the new features being introduced, rather than the whole product? If so, they’d be able to stop charging for iPod Touch updates, since they could still report the full revenue at the time of sale (whereas if they switched the iPod to subscription accounting today, revenue would suddenly drop to 1/8th what it is now).

All I can say is that it matches the quality of those on technical issues, marketing issues, and business too. Thank you Dan.

It also serves to prove just why I hate accountants and their blinkered view of the world.

This sounds like nonsense to me :

The only thing that Apple can fake when the phone is shipped and the money pocketed is development cost : that’s the only thing they are able to not deliver as promised (and they are able to not deliver it anyway because there’s no contract clause specifying when the upgrade come, with what, and how many bugs). On the other side, the subscriber is compelled to pay a precise amount.

I think it’s much more about administrative stupidity here on the Sarbanes-Oxley side.
On the other side, there’s a duty (or almost) for listed companies like Apple to smooth their income variation in order to play on Wall Street expectations, at least, not to get backlashes. A casual accounting trick is to provision income from the future to the present and reciprocally. It’s not illegal, and most of the times not even dishonest. Just the best angle to show the thing to other people. Why wouldn’t you spare $100.000 of profit to the R&D budget in this exceptionally lucrative n year and cut another $100.000 the bad n+1 year in R&D, compelling the R&D to dip into the first $100.000 to keep the budget even ?

It’s actually quite common.