waffle

The general impression of yesterday’s iPhone SDK for me is that Apple arguably over-delivered in terms of technical capability, but were right on the money in terms of the lock-in solution everyone figured they would use.

The APIs that are open are well thought-out and thoroughly documented. The tools are polished. A lot of useful sensor information is available. There’s nothing wrong with the capability of the SDK as far as an SDK goes; in terms of possibility per complexity (a unit I just now invented), it’s by far the best mobile API in the market.

Lock-in, however, is the big deal people still have problems with. And I don’t mean “writing apps just for the iPhone’s APIs” — almost every mobile cross-platform API sucks or is undergoing heavy development (hello, Android). If you were resigned to write iPhone apps using cross-platform or generic APIs, those would be the worst damn apps ever. No, I’m talking about the way Apple controls the development and distribution process.

Some decisions are understandable. If you use undocumented APIs, you won’t be able to get on the App Store, which seems like a bummer until you realize that undocumented APIs are highly fragile (or at least not sharing the explicit guarantee to be stable) — you want apps to be stable on your phone even if it’s got a stable UNIX-based OS underneath simply because you want them to keep working across firmware upgrades and not collapse into small ash piles. And I guess the “are you serious?” tax is a good idea for grooming the platform to some degree; to place the bar slightly above ground level, although many great apps are written by people who can’t easily reserve $99 to plunk down as a function of daily life (students).

However, the rest quickly reestablishes the image of (slick) interoperation at the cost of flexibility that Apple’s acquired over the years through constant repetition. Like yesterday, Apple is the sole distributor of “legitimate” (officially supported) iPhone applications. 30% of profits may be a reasonable limit in light of, say, the prospect of 40% of profits, but I’ll bet you euros to donuts that Apple makes up on everything involved to run the store after a measly 5-10%, and that the rest is mere gravy. The 30% being an iTunes parallel fails since there are no labels to please or MBAs (no, not MacBook Air, the other white meat) to invite around board room tables, and the developer tool funding is easily handled by the $99/$299 fee to get in under the tent.

Like Michael Tsai, I think that the decisions taken were based on sound thinking, but that the spin to convince the doubters leave Apple in an uncomfortable house of constructions: Not necessarily lies, but needless justifications serving only to strengthen a plan that already stands on its own.

I think the plan is well-executed, but I will continue to largely disagree with some of the actual decisions in it. There’s no “hacker’s switch” to port and run everything you ever wanted under the assumption that you’re fully responsible for anything you do from here on out. Well, there is, but like on Wednesday, it’s spelled “jailbreaking”. Within weeks or possibly days, the way Xcode gets apps on the phone will be hacked, and people will be able to do this for free and with unapproved apps. The same thing will happen for the App Store. And I’m left wondering why Apple didn’t just meet us halfway, since it is completely apparent to anyone in general and Steve Jobs in particular that this is going to happen.

The iPhone SDK isn’t just the best mobile SDK ever — and this isn’t saying much — but it’s also a great SDK, period. But like so many times with Apple, I wish they could have taken it a little bit further. (Actually, kind of like the Exchange ActiveSync reach-out.)