Paul Graham discusses Apple’s iPhone App Store approval process in his latest essay. I think it’s fair to summarize the core of his piece as:
- The App Store approval process is opaque and frustrating
- Its bad design is driven by AAPL’s perception that they own the channel between developers and users – AAPL is essentially trying to act as a “software publisher”
- 80’s companies like VisiCorp showed that software publishing makes no sense
- AAPL’s review process is also incompatible with launch-fast-and-iterate development, which is the way of the future
- By annoying its developers, AAPL is (to some extent) undermining its future
I think this essay overlooks some important things.
I’d argue that there are two big problems with this piece. First, AAPL isn’t wrong in their belief that they “own” their users; due to cryptographic techniques, they do own their users. Their position is completely different from that of 80’s “software publishers”, and more akin to that of, say, Nintendo. (Go look them up for an example of a difficult approval process.) Second, on the continuum of release rates, with cartridge-based, non-networked game consoles on one end (you’re never going to update anything) and web apps on the other (you can update several times a day, if you like) the iPhone, even with the bothersome App Store approval process, is closer to the Platonic ideal of launch-fast-and-iterate than desktop software (in practice) is today.
I wish the iPhone was completely open. I don’t like the idea of locked-down devices, or computers. I find them creepy. But pretending that the fact that a platform is locked down will spell its doom is ahistoric. If you want to deploy to the iPhone w/o hindrance from AAPL, build a web app targeted to Mobile Safari. That has problems too (it’s harder to charge money, and you don’t have access to all the capabilities of the device) but no one said life was perfect.