Conspiracy theories are almost always wrong. It certainly is the case here. Anyone who has worked on and shipped any serious piece of software should be able to tell you that this sort of mistake can and does happen. Testing can attempt to catch these problems, but doesn't guarantee it. Apple's testing emphasis was undoubtably the iPhone4 since it was new hardware, but they will also have had 3Gs and 3GSs in their testing matrix... but that doesn't guarantee that they would have seen the slow-down problems. Without knowing exactly what the problem(s) are, it is impossible to say how it was possible that their testing didn't catch the issue. It may also have manifested late in testing and not shown up as very bad so they decided to ship it with a plan to figure it out and fix it later... only it turned out to be much worse in the wild. And don't make the mistake of belittling them over it because it can and does happen to all developers. Problems in a company's development process show up in how often it happens, and how they deal with it when it does. Apple's track record in both areas is not perfect, but is generally pretty good.
Having a larger user base on the new OS is a good thing, the last thing they would want to do is sabotage it. A poor user experience is likely to drive users away from the platform, which is something they would not want to risk.
As for the expectation that successive OS releases must be slower, this is patently false. MacOSX in almost every version has actually gotten faster in most ways on the same hardware and, in particular, Leopard->SnowLeopard was notably a performance-oriented release. iOS3 is based on Leopard, and iOS4 is based on SnowLeopard. There are always trade-offs made for features vs performance, but many features can and are implement as zero cost if the application isn't actively using them -- they just consume a (small) amount of Flash space. Some of the iOS4 features clearly carry a performance and memory cost, but these should be balanced by other system optimizations. Apple doesn't want its iPhone4 to feel slower, it wants it to feel as fast as possible... and that will help the 3G and 3GS as well since they are fundamentally the same architecture and processor core, just with different clock rates and available memory. The problem on the 3G was very clearly a problem (i.e. a bug), as opposed to a set of design decisions made to cripple the older phones.
Providing grist for the rumour mill since 2001.
Providing grist for the rumour mill since 2001.
