There is still one point on which I kind of disagree. Strong typing makes the analysis easier, but I don’t think that it is the most difficult part. I rather beileve that the most difficult part comes from the fact that native code execution is arbitrary, and that it can be used to manipulate code pointers. Thismakes the analysis much more complex, because it greatly increases the possible number of execution patterns. There is already some work about such native code analysis, which works under the condition that the code satisfies some properties. This is quite likely to come in the near future.

If we come tu rumors, I would even say that static analysis is one good reason for Apple to force the use of their tools. If the code generated by their tools satisfies a few interesting properties, it makes advanced static analysis possible, which can significantly reduce the cost of vetting applications. Now that can be considered as a competitive advantage, or as a very nice way to balance automation and human discretion very efficiently.