Earlier quoted context omitted.
But, this isn't about giving the users freedom from choice; with 140,000 options in the App Store, that's a lot of choices. This is explicitly about locking developers down, and not allowing them to do things that Apple does not approve of. Want to write a competing web browser? Sorry, you can't. Want to write an awesome JIT that optimizes itself for their shiny new chip? No can do. Imagine if this policy had been in…
You could see it another way: Want to write a competing web browser that will confuse people (why do I need another web browser?), no can do. It's a quality control issue. The iPhone, iPod Touch, and iPad do specific things and do them well. They are not personal computers. They are web browsers, music players, calendars, etc. The customer is paying for a device that does these things. They don't want (or arguably ne…
And Apple is not immune to this sort of problem; there are alternative products that they forbid that really can improve the experience (for instance, better podcast clients, email clients that are optimized for Gmail, and the like). And given the quantity of software they have on the app store, and how crappy some of it is, it's not about quality control at all; it's about the fact that Apple simply wants control, and doesn't want any third-party developers getting too big or too powerful.
The point about LLVM is not about developing LLVM on an iPad. It's about the fact that you can't use such technology on the iPad. If I want to develop a product to run on the iPad, and that product would be greatly improved by containing a JIT or an interpreter, I cannot distribute that on the iPad; it is forbidden. And this means that there is less of a market for developing such things in general. If the hardware that the consumers have and use forbids me from selling them a product with an interpreter in it, then there's a lot less of a reason for me to produce that interpreter in the first place.
I write interactive educational multimedia software that's scripted in a dialect of Scheme. The iPad may be a great platform that we could port our software to; the screen is a perfect size for our type of content, it supports video, composited graphics and animation, and multitouch interaction could really improve some of our simulations. However, we cannot run our Scheme interpreter on the iPad. Writing a compiler would be prohibitively expensive for our small team, as would porting all of our content to JavaScript, the only interpreted language you are allowed to run on the iPhone/iPad. Apple is explicitly preventing me from porting the software I would like to the iPad.
But you don't hear as much about people like me, who simply turn away from the iPhone or iPad and don't develop software for it. You hear about the success stories; the people who made a quick buck selling a fart app for the iPhone. The only cases in which you do hear about people being turned away are cases in which people did make the initial investment, developed the software, and then were turned down; but because most people don't want to risk that, they never bother developing the software in the first place if they think there's a high chance that Apple will reject it.