On the damn-the-torpedoes side, there's a fairly irate post by Hadley Stern at AppleMatters calling Apple's decision not to release a full SDK "stupid and arrogant," and arguing that the incredible potential of open development for the iPhone far outweighs any security considerations for the device. Even though the post is billed as "10 reasons to open iPhone" and by my count there are only 7 listed, there's little doubt that the imaginations and energy of independent developers would create wonders on the iPhone platform -- taking note of what's already been accomplished. Hadley's crankiness notwithstanding, the post raises interesting parallels between iPhone and previous Apple 'closed' platforms, including one spectacular success despite a private API (the iPod).
On the watchful-waiting side, there's our colleague Alex Hung over at Download Squad, with a more thoughtful piece on the reasons why an iPhone SDK isn't ready for prime time yet. While a fully functional toolset with Apple's documentation and support is both a good idea and seemingly inevitable, Alex points out that locking down the APIs at this stage of the product's lifecycle would box in the engineers who are still working to make the iPhone everything it can be. Any iPhone SDK would also be (most likely) tightly linked to the Leopard code base, so trying to regress the SDK to Tiger while simultaneously developing for OS X iPhone's future would be a useless exercise in platespinning.
I'd put myself pretty firmly in the latter camp. While the work of the #iphone folk in compiling a working development environment for the iPhone is both astonishing and commendable, I can't see the point in opening the iPhone to official 3rd-party development with a fully-blessed SDK until everything is completely baked. On the heels of the Leopard release, with an eager installed base of devices and a dev community primed by the unofficial tools to make the best of it, a true SDK for iPhone -- ready for prime time -- will be a Good Thing.