The idea that some molecules must have been designed, while other "obviously" are not, is a crock of s**t.
The point I made is that your example of the order of water molecules doesn't rise to the complexity of a machine with irreducible parts.
First, you can remove any number of the molecules without reducing the function of the machine. Thus, your ice machine has no irreducible core.
Second, the arrangement of your machine is not complex; it is the perfect, natural result of water getting cold. It doesn't begin to compare with the complexity of any machine.
Third, the function of your machine is rather accidental and happenstance. That is not to say that a block of ice laying across a ditch couldn't be used to walk across, but it is saying that the function has nothing to do with the reason the block exists as it does. This sets up nicely your claim that without a designer to call it functional, you have no function, and thus without an appeal to the designer, you have no IC or ID. The problem, though, is that in a machine such as flagellum, it doesn't matter who the observer is--the machine does work. It has a function independant of an observer coming along and walking across it. And further, whether flagellum evolved or was designed, it has a specific function that justifies its existence. Your block of ice can only claim this if something with intelligence comes along after the fact and invents it.
When you say "the idea that some molecules must have been designed, while others 'obviously' are not, is a crock..." you are correct. That is why no one on the ID side says it, and why your reference to complex clays is a strawman.
Mistook you for MHal; didn't notice a different person was answering, thus there are numerous "yous" that refer to someone else. You shouldn't mind, though, because it seems you were picking up where MHal left off.